Antigen-specific activation of memory CD8+ T cells

Eur. J. Lmmunol. 1992. 22: 1595-1601

Ferdynand J. Kos and Amo Miillbacher Division of Cell Biology, John Curtin School of Medical Research, Australian National University, Canberra

1595

Specific epitope-induced conversion of CDS+ memory cells into effector cytotoxic T lymphocytes in vitro: presentation of peptide antigen by CDS+ Tcells The requirements for the conversion of CD8+ memory Tcells into effector class I major histocompatibility complex (MHC) Kd-restricted cytotoxic T (T,) cells in vitro have been studied. Purified CD8+ splenocytes from influenza A/WSNprimed BALB/c (H-2d) mice stimulated with a synthetic nucleoprotein peptide 147-158 R156- (NPP) alone generated T, cells specific for influenza virusinfected target cells. No additional requirements for accessory cells or their lymphokine products were necessary indicating that peptide antigen (Ag) in association with Kd waspresented on CD8+T cells.The evidence for presentation of NPP by CD8+ T cells was supported by the use of CD8+ memory T cells from semiallogeneic bone marrow radiation chimeras of PI’ F1 type (H-2b’ [H-2d x H-2b]F1). Memory CD8+ splenocytes from A/WSN-immune chimeras did not develop into secondary effector T, cells as a result of a 4-day culture with NPP alone, however, were able to do so if NPP was presented by Kd-bearing Ag-presenting cells. In addition, these results exclude the possibility of direct recognition of free NPP molecules by the specificT cell receptor of CD8+ memory Tcells. CD8+ memory splenocytes (H-2b) from chimeras were also able to develop into functionally activeT, cells as a result of presentation of Dh-restricted synthetic peptide (NP 366-374) with a sequence derived from influenza virus nucleoprotein with high affinity for Dh MHC class I molecules. Blockade of endogenously produced interleukin 2 (IL-2) activity by anti-IL-2 or anti-IL-2 receptor monoclonal antibody in the culture of CD8+ memory Tcells during a 4-day NPP stimulation completely abolished T, cell generation, indicating that the utilization of this lymphokine is absolutely required for the secondary T, cell development. These findings demonstrate that CD8+ memory T cells per se are able to recognize the restimulating epitope as a result of its presentation by CD8+ Tcells and develop into cytolytically active and highly specific T, cells with no requirements for other cellular helper components or their lymphokine products.

1 Introduction Influenza virus infection induces in mice strong cytotoxicT (T,) cell responses against most viral components (reviewed in [l]).The majority of MHC class I restricted T, cell clones are specific for the influenza nucleoprotein (NP) [2-41. Dominant influenza NP determinants are located in amino acid residues 147-158 for Kd-restricted T, cells [ S ] , residues 365-380 for Dh-restricted T, cells [6], and residues 50-63 for Kk-restricted T, cells [7]. It has been suggested that effector T, cells are not necessarily terminal cells but may revert, at least in vitro, to memory T, cells [8]. Ag-specific memory T cells express enhanced responsiveness to recall Ag and may respond to Ag qualitatively and/or quantitatively different than do naiveT cells [9-111. The requirements for the conversion of memory T, cells into effector Ag-specific T, cells are poorly defined despite the progress in the characterization of phenotypic and functional properties of

[I 103031 Correspondence: Arno Mullbacher, Australian National University, John Curtin School of Medical Research, Division o f Cell Biology, PO. Box 334, Canberra, ACT 2601, Australia Abbreviations: AIWSN: Influenza virus strain AIWSN133 (HlN,) NP: Nucleoprotein NPP: Synthetic NP peptide 147158 R- Tc: Cytotoxic T (cell)

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1992

memory Tcells [lo, 111. The main reason for our lack of understanding of T, cell activation is due to the complexity which arises from the following: (a) a very low frequency of clonally distributed Ag-specific T cells in the population, representing only tenths to hundredths of a percent of unselected T cells [ 121, (b) impossibility to separate epitope-specific T, cells from bystander T, cells of irrelevant specificity (Ag-specific T cell clones of uniform specificity are maintained in an established effector phase of their development and are not suitable for studying the activation requirements), and (c) complex intermolecular interactions that occur at the level of Ag presentation and signalling by the APC. Exploring this issue we have attempted to circumvent some of the problems mentioned above by adapting a system for nucleoprotein (NP)-specific Tcell clones, originally described by Townsend et al. [4-6, 131, to study the conditions for the conversion of murine CD8+ memory Tcells into effector MHC class I (Kd)restricted T, cells specific for influenza A virus-infected cells in vitro. We demonstrate here, that highly purified CDX+ influenza-immune splenocytes stimulated with a synthetic NP peptide (NPP) 147-158 R 156- alone respond to NPP-Kd with the generation of cytolytically active effector T, cells revealing thereby the minimal requirements for this process, namely presentation of NPP by CD8+ Tcells without the need for accessory cells or lymphokines provided by other leukocytes. 0014-2980/92/0606-1595$3.50+ ,2510

1596

F. J. Kos and A. Mullbacher

2 Materials and methods

Em. J. Immunol. 1992. 22: 1595-1601

2.1 Mice

anti-mouse dendritic cell mAb 33D1 [American Type Culture Collection (ATCC), Rockville, MD]. The efficacy of cell depletion was determined by flow cytometry analysis.

Female BALB/c (H-2d), C57BL/6J (H-2b), CBA/H (H-2k) and (BALB/c x C57BL/6J)F1 hybrid (H-2dxb)mice were obtained from the Animal Breeding Establishment of the John Curtin School of Medical Research. All mice used in these experiments were 6-12 weeks of age and were maintained under specific pathogen-free conditions.

2.5 Separation of CDS+ Tcell population by FCM

2.2 Cell culture P815 (H-2d)mastocytoma, EL4 (H-2b)thymoma, and L929 (H-2k) fibroblastoma cell lines were maintained in culture in Eagle's minimum essential medium with Earle's salts (EMEM; Gibco, Grand Island, NY, USA) supplemented with 5% heat-inactivated fetal bovine serum (FBS; Pacific Bioindustries, Sydney, NSW, Australia) and antibiotics (penicillin 100U/ml and streptomycin 100 yg/ml) at 37 "C in a humidified atmosphere of 5% COz in air. Isolated spleen lymphocytes were cultured in EMEM with 10% heatinactivated FBS, 5 X lop5 M 2-ME, polymyxin B (Sigma, St. Louis, MO, 10 pg/ml), penicillin, and streptomycin, hereafter referred to as the culture medium.

2.3 Antigens Influenza virus strain A/WSN/33 (HIN1) (A/WSN) was grown and titrated as previously described [14]. A/WSN was used in the form of infectious allantoic fluid. Kdrestricted NPP was synthesized as a modified NP sequence 147-158 R156- [13] comprising 11 amino acids (TYQRTRALVTG) by Peptech Co., Sydney, Australia, in accordance with the method described by Townsend et al. [6]. Db-restricted NP sequence 366-374 (ASNENMETM) [151 was synthesized by Biomolecular Resource Facility (Australian National University, Canberra, Australia). This peptide was synthesized by the solid-phase method carried out on the Applied Biosystems (Warrington, GB) Model 430A peptide synthesizer. After purification by HPLC, the peptide assembly was verified by amino acid analysis.

2.4 Preparation of enriched CDS+ T cell populations Ficoll-passed splenocytes from mice primed i.v. with A/WSN [1 X lo3 hemagglutination units (HAU) per mouse] 3-5 weeks previously,were depleted of Ig+ cells by repeated (2 X ) rosetting with sheep anti-mouse Ig-coupled immunomagnetic beads; Dynabeads M-450 Tosylactivated (Dynal, Oslo, Norway) were covalently bound with antimouse Ig mAb (Silenus Laboratories, Hawthorn, VIC, Australia) according to a procedure provided by Dynal. Igsplenocytes were depleted of CD4+, asialo-GMl+, 33Dl+ cells, and in some control experiments Thy-l+ or CD8+ cells, by repeated treatment with mAb plus complement (C; Cedarlane Laboratories, Hornby, Ontario, Canada). The mAb were as follows: anti-mouse Thy-1.2 clone F7D5 (Serotec, Kidlington, GB), anti-mouse CD4 clone RL174 and anti-mouse CD8 clone 31m (both provided by Dr. R. Ceredig LGME, Strasbourg, France), anti-asialo-GM1 (Wako Pure Chemical Industries, Osaka, Japan), and

Samples of CD8+-enriched T cells (= 95% of CD8+), after treatment of splenocytes with a coctail of mAb plus C, as described above, were stained (30 min at 4°C) with a predetermined optimal amount of anti-mouse Ly-2 FITCconjugated mAb (Becton Dickinson, San Jose, CA). Fluorescent samples were sorted on a FACS 440 flow cytometer (Becton Dickinson) at 4°C using an argon ion laser tuned to 488 nm and running at 200 mW. Fluorescence was detected through a 535/15 bp filter and CD8+ (Ly-2+) cells were collected on the basis of a combination of narrow-angle forward scatter and fluorescence intensity. Small samples of sorted cells were reanalyzed to determine their purity.

2.6 Generation of effector T, cells from memory CDS+ T, cells Purified CD8+ T cells from A/WSN-immune mice were stimulated in triplicate cultures in Linbro 96-well roundbottom plates (Flow Laboratories, McLean,VA) at 5 x lo5 cells/well/0.2 ml with NPP at 1 0 - l O ~After . 4 days of incubation the effector cells of each individual culture well were fractionated into 3-fold diluted aliquots and assayed in a 4 h SICr-release cytotoxicity assay against 51Cr-labeled (Amersham, GB) uninfected P815 (P815 [NIL]) or A/WSN-infected (700 HAW1 x 106-3 x lo6cells/0.2 ml for 1 h at 37°C) P815 (P815 [A/WSN]) target cells (2 x 103/well/0.2ml). The cytotoxicity assay was performed according to standard procedure as described previously [14,16]. Results are expressed as the percentage of specific lysis at the determined fraction of original culture assayed. Specific lysis (SL) was determined as: YO SL = 100 x [(experimental release - spontaneous release)/(maximum release - spontaneous release)]. Spontaneous release was always < 10% of the maximal release. For co-culture experiments, titrated numbers of CD8+ T cells were co-cultured with fixed numbers of CD4+ T cells for 4 days in the presence of NPP M) and the activity of generated effector Tc cells was assessed in the cytotoxicity assay against P815 [A/WSN] target cells. In some experiments, aliquots of effector cells after 5 days of culture were treated with: (1) C alone, (2) C plus anti-Thy-1.2, (3) C plus anti-CD4, (4) C plus anti-CD8, or (5) C plus anti-asialo-GMl, as described above. These treated cells were then tested in the cytotoxicity assay. 2.7 Establishment of semiallogeneic bone marrow radiation chimeras

Recipient (BALB/c X C57BL/6J)F1 hybrid mice were lethally irradiated with 950 rad from a mCo source 24 h before intravenous transfer of 15 x lo6 anti-Thy-1.2 plus C-treated bone marrow cells from C57BL/6J donor mice.

Antigen-specific activation of memory CD8+ T cells

Eur. J. Immunol. 1992. 22: 1595-1601

The chimeras were given an antibiotic terramycin (1 g/l; Pfizer Agricare Pty. Ltd., West Ryde, N.S.W., Australia) in drinking water and rested for 12 weeks. The animals were subsequently immunized with influenza virus A/WSN and after 3 weeks tested in experiments described below.

1597

i 0 .'c

u

Q Q

2.8 Generation of alloreactive T, cells in mixed lymphocyte culture (MLC) experiments in vitro Spleen stimulator cells were prepared by treatment with mitomycin c (Sigma) at a concentration of 50 yg/ml/2 X lo7 cells for 1 h at 37 "C followed by four washes with EMEM. Two million spleen responder cells were co-cultured with 1 x lo6 stimulators in 2 ml of culture medium in Linbro 24-well tissue culture plates (Flow Laboratories) for 5 days at 37°C in a humidified atmosphere of 5% C02 in air. Cytotoxicity assay on 51Cr-labeled EL4 and L929 target cells was performed as described above. 2.9 Antibodies used for blockade of T, cell generation Antibody neutralizing murine IL-2, produced by S4B6 hybridoma cells and used in the form of ascites fluid, was kindly provided by Dr. €? Hodgkin (DNAX, Palo Alto, CA). Hybridoma PC 61.5.3, secreting anti-mouse IL-2R mAb, was obtained from the ATCC; this mAb was prepared as culture supernatant of cells grown to saturation in DMEM supplemented with FBS (5%), glucose (4.5 g/l), 2-ME (2 X lop5M), and antibiotics.

v)

.I

.01

Culture

Figure 1 . Generation of secondary anti-AIWSN T, cells by in vitro stimulation of purified CD8+ T cells from A/WSN-immune mice with peptide Ag. CD8+ Tcells were isolated from BALBIc mice as described in Sect. 2.5 and cultured (5 X lo5 cellslwell) for 4 days in the presence of NPP (lo-'" M). O n day 5, the effector cells were assayed for cytolytic activity against 51Cr-labeleduninfected (0)or A/WSN-infected (0)P815 target cells. Results are expressed as mean percentage of specific lysis SE at the determined fraction of original culture assayed.

+

A.

.-

v)

* v)

-I

.-r 0

P aJ

fn

d 7 0 I$o 4

3 Results

1

Fraction

I

o5 CDB+ Cell No./Well

B. I

3.1 Minimal requirements for the activation of CDS+ memory T cells by NPP Purified influenza-immune memory CD8+ splenocytes were stimulated with a synthetic peptide NPP, with a sequence derived from influenza virus NP with high affinity for Kd MHC class I molecules [13], to generate cytolytically active effector cells specific for A/WSN-infected target cells in vitro (Fig. 1). Highly purified CD8+ Tcells (>98% of Ly-2+ cells, as determined by flow cytometry analysis) stimulated with NPP responded with the generation of secondary T, cells cytolytically active against A/WSNinfected P815 target cells. No further exogenous mediators were required indicating that all necessary signals for the conversion of CD8+ memory Tcells into effector T, cells can be provided by CD8+ T cells themselves.

To optimize the conditions in our model system we performed a series of titration experiments and their results are presented in Fig. 2. Measurable cytolytic activity of the generated secondary T, cells, expressed as the percentage of specific lysis of P815 [A/WSN] target cells, was observed in the range of CD8+ cell numbers between 0.1 x lo6 and 1 x lo6 per 0.2 ml culture well at the onset of the culture (Fig. 2 a). The establishment of a dose-response curve showed that the generation of effector T, cells was significant at concentrations of NPP L M (Fig. 2b). To determine the kinetics of the generation of secondary T, cells we measured lytic activity as a function of time after addition of NPP to CD8+ Tcells from A/WSN-immune

z 1

2

3

4

5

Time of Stimulation with NPP [days]

Figure 2. Response of purified CD8+ Tcells from influenza AIWSN-immune BALBIc mice to NPP stimulation expressed as cytolytic activity of in vitro generated T, cells. Percentage of specific lysis of AIWSN-infected (0) and uninfected (0) PS15 target cells is shown as a function of CD8+ cell numbers per culture well (A), NPP concentration (B), and time of stimulation (C).The number of CD8+ Tcells per well was 1 x loh for (B) and (C). NPP concentration used for stimulation of CD8+ memory Tcells in (A) and (C) was lo-"' M, and the time of stimulation for (A) and (B) 4 days. Results are expressed as mean values f SE of triplicate cultures.

1598

F. J. Kos and A. Miillbacher

Eur. J. Immunol. 1992. 22: 1595-1601

mice. As shown in Fig. 2 c, the maximal cytolytic activity was achieved when the cultures of CD8+ Tcells were stimulated for 4 to 5 days. The conditions chosen for all other experiments, unless otherwise stated, were as follows: (1) 0.5 x 106 CD8+ Tcells/well/0.2 ml, (2) NPP concentration M, and (3) stimulation time 4-5 days.

molecules on the surface of CD8+ Tcells is responsible for the conversion of memory into effector T, cells. To test this speculation we used semiallogeneic bone marrow radiation chimeras constructed by the repopulation of (H-2d x H-2b)F1 hybrids with H-2b bone marrow cells. The reason for their use was that these chimeras, when primed with A/WSN influenza virus, should generate virus-specific T, A/WSN-immune effector T, cells generated as a result of cells of H-2b genotype restricted to the host H-2dxb.All 5-day culture with NPP retained their CD8+ phenotype. chimeras were phenotyped, to evaluate the efficiency of Depletion of effector cell suspensions with specific mAb reconstitution, by flow cytometry analysis of splenocytes. plus C indicates that neither anti-CD4 nor anti-asialo-GM1 Splenocytes from all the chimeras tested stained with treatment of effector cells caused loss of lysis of A/WSN- anti-Kb mAb but not with anti-Kd mAb which in turn infected P815 target cells. In contrast, effector cells were efficiently labeled cells from control F1 animals (Table 2). totally sensitive to anti-Thy-1.2 or antLCD8 plus C treat- Thus, all of these chimeras were fully reconstituted by ment. Thus the effector cells mediating lysis were conven- donor-type (H-2b) hemopoietic cells. CD8+ A/WSNtional Thy-1.2+, CD8+, CD4-, and asialo-GM1- T, cells immuneT cells from chimeras, cultured for 4 days with NPP (Table 1). did not generate effectors which could lyse Kd-bearingP815 [A/WSN] target cells (Tables 2 and 3). However, the same To exclude further a role for CD4+ T cells in the activation CD8+ T memory precursor cells from A/WSN-primed of CD8+ memory Tcells via a small contaminating MHC chimeras developed into effector T, cells when stimulaclass 11-positive APC population, presenting either the ted with NPP-modified mitomycin C-treated P815 cells specific peptide or nonspecific FBS-derived peptides the (Table 3), revealing that the TcR of H-2b-bearing CD8+ following experiment was carried out. Graded numbers of memory Tcells could recognize NPP if presented by purified CD8+ Tcells (these cells contained < 1.6% CD4+ Kd-bearing APC. These results provide evidence that cells) were stimulated with NPP in culture together with priming of chimeras with A/WSN virus generated CD8+ T, fixed numbers of CD4+ Tcells (these cells contained cells with TcR specific for NPP-Kd. Moreover, donor < 0.3% CD8+ cells). After 4 days the cytolytic activity was H-2b-bearing CD8+ T cells from A/WSN-immune chimeras measured in a 4 h 51Cr-releaseassay against P815 [A/WSN] were able to present the Db-restricted synthetic peptide target cells. As shown in Fig. 3, CD4+ Tcells caused only a (NP 366-374) resulting in the generation of Db-restricted T, marginal increase in reactivity of the generated effector T, cells (Table 2). In addition, their ability to generate effector cells and only when the numbers of CD4+ T cells added to T, cells could be demonstrated by using mixed lymphocyte the suboptimal numbers of CD8+ cells (< 1.25 x 105/well) culture experiments to third-party alloantigens. CD8+ T cells from chimeras could differentiate into effector T, were high, that is, at CD4+ : CD8+ ratio above 1 : 1. cells albeit with lower lytic activity than CD8+ Tcells of donor or recipient origin (Table 2). All these results clearly suggest that the conversion of CD8+ memory Tcells into 3.2 Presentation of NPP by Kd molecules on CD8+ effector T, cells is a consequence of presentation of peptide T cells It is generally accepted that CD8+ T cells recognize an epitope formed by the association of a peptide Ag fragment with a MHC class I molecule on APC or target cells. In our system we postulate that presentation of NPP by Kd

; u)

Table 1. Phenotype of in vitro generated secondary A/WSNimmune T, cells as determined by mAb plus C treatment

4o

* NIL * 250,000 _Y

,......,,*

125,000 62,500

0

0

g Treatmenta)

Percent specific lysis at E :T ratio 6 : 1 P815 P815 AWSN] "ILI

None C alone Anti-Thy-1.2 C Anti-CD4 + C Anti-asialo-GMI + C Anti-CD8 C

+

+

41.07 40.11 - 1.28

1.52 - 1.27 - 2.34

40.18

0.09

38.30 - 0.36

- 0.88 - 0.23

a) Isolated CD8+ cells from A/WSN-primed BALB/c mice were cultured for 5 days with NPP (lo-'" M). Cells were then harvested, counted, divided into 6 aliquots and treated with mAb plus C as indicated. Specific lysis was tested in 4 h 51Cr-release cytotoxicity assay against A/WSN-infected or uninfected PS15 target cells.

20

u)

8

10

0

NIL

I o4

I

o5

106

CD8+ Cell No./Well

Figure 3. Generation of effector T, cells from memory CD8+ T, cells in the presence of CD4+ Tcells. Titrated numbers of purified CD8+ cells from AIWSN-immune mice were cultured with NPP (10-I" M) in the presence of indicated fixed numbers of CD4+ cells. After 4 days, the cells were spun down, the culture medium replaced by one containing "Cr-labeled P815 [A/WSN] target cells (1 X 104/we11/0.2mi), and cytolytic activity of generated effectors measured in 4 h SICr-release assay. Numbers given on X-axis represent the number of CD8+ Tcells per well at the initiation of culture. Data points show the mean values k SE of triplicate cultures.

Antigen-specific activation of memory CDW T cclls

Eur. J. Immunol. 1992. 22: 1595-1601

A.

1599

B. 1

60 1

I

60 1

50

40

30 20

10 0

1

1 0

100

1000

10000

Reciprocal of Dilution of Anti-IL2R

NIL

1

mAb

1 0

100

1000

10000

NIL

Reciprocal of Dilution of Anti-IL2 mAb

Figure 4. Requirement for endogenous IL-2 activity in the conversion of CD8+ memory T, cells into effector T, cells in vitro. A/WSN-immuneT, cells were generated as in Fig. 1but in the presence of anti-IL-2R (A) or anti-IL-2 (B) mAb.The number of CD8+ cells per well was 5 X lo5.Cytolytic activity was assessed in a 4 h 5LCr-releaseassay on P815 [A/WSN] target cells.The data are expressed as the mean percentage of specific lysis f SE of triplicate cultures.

Table 2. Phenotypic and functional properties of splenocytes from semiallogeneic bone marrow radiation chimeras

Source

Responding CD8+ T cellsa) Surface expression of Kd and Kb moleculesb) Kd

Kb

Stimulation

NPPC)

NP CBAN (366-374)c) (Mit. C)d) % Specific lysis of P815 EL4 L929 (ATWSN) (ATWSN)

a) Ficoll-passed splenocytes from A/WSN-primed mice were used for preparation of enriched CD8+ 27.95 1.34 39.30 Chimera No. 1 Tcell population, as described in 31.05 0.38 28.35 Chimera No. 2 Sect. 2.4. and then cultured for 5 1.95 40.97 28.84 Chimera No. 3 days with an indicated stimulus. 25.07 2.05 34.95 Chimera No. 4 The activity of effector cells was 25.02 0.29 42.58 Chimera No. 5 assayed in a 4-h 51Cr-releasecyto27.15 6.62 30.47 Chimera No. 6 toxicity assay against 51Cr-labeled 42.45 48.12 55.36 [BALBk X C57BL/6J]F1 target cells (PSI5 and EL4 target 45.63 0.75 57.09 BALBIc cells were additionally infected 40.73 56.10 4.89 C57BLl6J with AIWSN) as described in Sect. 2.6. b) Flow cytometric analysis of the distribution of Kd and Kh molecules on the surface of responding cells was performed using FACScan (Becton Dickinson). The cells were labeled with anti-Kd and anti-Kh mAb (used as a culture supernatant of hybridoma cell lines SF1-1.1.1 and 28-13-38, respectively; ATCC and stained with FITC-conjugated sheep anti-mouse Ig mAb. c) Both NPP and NP (366-374) were added to the cultures of responding cells (1 x 106/we11/0.2ml in 96-well plates) at the concentration of 1 0 - 6 M . d) Mitomycin C-treated stimulating CBAlH spleen cells were used for the generation of alloreactive T, cells in MLC, as described in Sect. 2.8.

Table 3. Requirement for Kd-bearingstimulator cells for the in vitro generation of NPP-Kd-specific effector T, cells from H-2h+ (H-2" x H-2h)F1 semiallogeneic bone marrow chimeras

In vitro stimulation

Responding splenocytes Surface expression of Kd and Kb moleculesb)

of mice primed with

Kd

Kb

Chimera No. 7 Chimera No. 8

-

BALB/c

+

+ +-

a) Ficoll-passed splcnocytes from mice primed i.v. with AlWSN 3 weeks previously were cultured for 5 days at the concentration of 4 x 10h/well/2ml with either NPPalone (1W6 M) or NPP-labeled M, 1 h, 37°C) and mitornycin Ctreated (50 pg/ml, 1 h, 37°C) P815 stimulator cclls (1 x 1Oh/well/2ml) in 24-well plates.

NPP

P815 (NPP)Mit.c % Specific lysis OF) P815 P815 P815 P815 (AIWSN) (NIL) (ATWSN) (NIL) 0.79 3.90 27.26

0.29 0.36 1.61

21.64 20.93 38.38

0.77 1.30 3.34

b) See footnotes to the Table 2. c) T, cell response assayed in 4-h sLCr-releasecytotoxicity assay against indicated targets (1 percentage of specific lysis at the culture fraction 0.1.

X

104/well/0.2ml) is expressed as the mean

1600

F. J. Kos and A. Mullbacher

Ag within the population of CD8+ Tcells. At the same time, these results exclude the possibility of direct recognition of free NPP molecules (i.e., not presented by MHC molecules) by the TcR of CD8+ memory T cells.

3.3 Requirement for endogenous IL-2 activity in the generation of secondary T, cells The role of IL-2 in the maturation of CD8+ memory T cells into effector T, cells in vitro was assessed. The addition of anti-IL-2R or anti-IL-2 mAb at the commencement of cultures of purified CD8+ Tcells totally abolished the generation of effector T, cells cytolytically active against P815 [A/WSN] target cells (Fig. 4). This observation suggests that IL-2 produced endogenously by CD8+ T cells (data not shown) is absolutely required for the generation of experimentally demonstrable secondary T, cell responses in vitro.

4 Discussion In this report we have examined the requirements for the induction of secondary T, cell responses using purified CD8+ Tcell populations. Splenocytes primed in vivo with influenza A/WSN virus and restimulated in vitro with the synthetic NPP peptide generated T, cells specific for A/WSN-infected target cells. The process of the conversion of NPP-Kd-responding T cell precursors (memory T cells) into effector T, cells occurred efficiently under in vitro conditions with no obvious requirements for accessory cells or their lymphokine products. This demonstrates the presentation of the NPP-Kd epitope by CD8+ Tcells followed by acquisition of cytolytic activity by effector T, cells. T, cell activation is a complex phenomenon and it is tought to involve collaboration of several cell populations. The prevailing opinion states that the induction of most functions of CD8+ Tcells requires help provided by CD4+ Tcells and/or conventional APC [17, 181. However, there exists evidence that T, cells can be generated in the absence of CD4+ Tcells [19-241 and that CD8+ Tcells represent a heterogenous population of collaborating subsets [25, 261. In addition, the requirements for the stimulation of memory Tcells are less stringent than those for naive resting T cells [ 101.The phenotype of memory T cells, used in various functional studies, can partially be distinguished from naive T cells by unique surface molecules [lo, 271. However, this approach is unable to separate Ag-specific T cells from bystander T cells of irrelevant specificity. Here we use a functional distinction of specific memory Tcells as defined by the process of recall of epitope-specific T, cell precursors to study the activation requirements. Vitiello et al. [28] in their studies on the consequences of the in vitro interaction of MHC class I-restricted T, cell clones with soluble Ag demonstrated the phenomenon of self-presentation of peptide by MHC class I molecules on the effector cells. We demonstrate here that populations of purified CD8+ T cells from immune mice are able to present peptide Ag in association with MHC class I molecules to CD8+ T cell precursors and induce their conversion into

Eur. J. Immunol. 1992. 22: 1595-1601

cytolytically active secondary effector T, cells in vitro. This process occurs efficiently despite the relatively low frequency of NPP-Kd-respondingT, cell precursors estimated at 1 in 85 000 CD8+ T cells, as measured in a filler cell-free limiting dilution assay (unpublished observation). It has already been shown that many different cell types have the capacity to function as APC for unprimed CD8+ T cells in the absence of added help [29, 301. Our observation extends these results by providing the evidence that CD8+ Tcells, purified from immune mice, can efficiently act as APC for memory CD8+ Tcells in vitro. The role of IL-2 in the growth of Tcells is well documented and our observation that IL-2 endogenously produced by restimulated CD8+ T cells was both necessary and sufficient for T, cell generation is not surprising (see also [25,31-331). The behavior of memory CD8+ Tcells in this respect resembles well-characterized helper-independent T, cell clones producing their own IL-2 acting as an autocrine helper factor [19].There are two possibilities by which IL-2 may act in our system: (a) IL-2 may be necessary for the Ag-specific activation of memory CD8+ T, cell precursors and/or (b) IL-2 may be required for the proliferation of activated memory/effector CD8+ T, cells and their clonal expansion results in the generation of detectable cytolytic activity in an in vitro assay. Recent study by Horak et al. [34] suggests that both activities should be considered in future experiments on the role of IL-2 in the Ag-specificactivation of T cells. They have shown that IL-2 transiently stimulates the specific activity of p561Ck,a nonreceptor tyrosine protein kinase, participating in signal transduction in IL-2-dependent Tcell clones by the formation of the complex with either CD4 or CD8. The authors postulated further that IL-2 may play a dual role in Ag-dependent T cell activation events and in the proliferation pathways [34]. The existing theoretical models of self-nonself discrimination and T cell activation, especially the associative recognition theory of Bretscher and Cohn [35] and Langman [36], assume that two Ag-specific signals are involved in the activation of T cells. More recently, a model resolving some of the paradoxes of the above-mentioned theory was proposed by Liu and Mullbacher [37]. They speculate that immunological help is reciprocally delivered between different lymphocyte subpopulations. For a lymphocyte (Tor B) to become fully functionally activated two signals are required; signal 1is provided by Agper se and signal 2 by an Ag-stimulated immunocompetent cell. In the case of T cell-T cell collaboration, in its minimal model, an additional cell expressing MHC class I1 molecules has to be involved [37]. Here we describe a situation where Ag alone (single epitope recognized in MHC class I-restricted manner) is sufficient to induce activation of Ag-specificmemory T, cells in a population of CD8+ Tcells purified from influenza-immune mice. One of the prerequisites would be that collaborating CD8+ T cells are functionally different, e.g. secrete different lymphokine(s). In our model, we assume no participation of MHC class 11-positive cells in this process. However, the recent findings by Perkins et al. [38] that NPP is also immunogenic in the context of I-Ad molecules raise some doubts regarding this assumption. Thus, even in this minimal antigenic requirement model (a peptide of 11 amino acids) the possibility that more than one T cell epitope is available cannot be ruled out. Therefore, we can not formally exclude that the two-signal model

Eur. J. Immunol. 1992. 22: 1595-1601

of Tcell activation is aplicable for memory T, cell activation. A physiological rationale of a situation, where CD8+ T cells can be recognized by CD8+ T, cells seems to be obvious. CD8+ cell displaying nonself Ag in the context of MHC class I molecules should be recognizable by MHC class Irestricted T, cells irrespective of the differentiation stage of the T, cell (naive precursor, memory, or effector cell). The stage of differentiation of the particular cell type which is to be activated is critical. According to Langman, “new” (resting) Ag-reactive cells require two signals to be activated whereas the effectors derived from these same cells can interact with Ag, expressing their effector function without the need for signal 2 [36].What is the nature of memory CD8+ T, cell in this respect we do not know. Are they quiescent resting cells or quiescent effector cells? Do they need a qualitatively different Ag-specific signal 2 for their activation? Perhaps, memory CD8+ T, cells need only signal 1 and auto- or paracrine action of lymphokines (e.g. IL-2) provides help for their activation, and other mechanisms (e.g., apoptosis or self-lysis of effectors) ensure the elimination of the risk of autoimmunity by development of mutated self-reactive CD8+ T, cells. In conclusion, we found that purified CD8+ Tcells from immune mice are able to recognize the restimulating epitope on CD8+ Tcells and undergo functional maturation into effector T, cells providing all the necessary factors (e.g. IL-2) without any further requirements. These data support the concept that individual CD8+T cells have much more functional flexibility than previously suspected. We would like to thank Drs. Robert Blanden, Markus Simon, and Gordon Ada for critical reading of this manuscript and helpful discussions.

Received January 24, 1992; in revised form March 9, 1992.

5 References 1 Yewdell, J. W. and Hackett, C. J., in Krug, R. M. (Ed.), The Influenza Viruses, Plenum Press, New York 1989, p. 361. 2 Kees, U. and Krammer, F! H., J. Exp. Med. 1984. 159: 365. 3 Townsend, A. R. M. and Skehel, J. J., J. Exp. Med. 1984.160: 552. 4 Townsend, A . R. M., Gotch, F. M. and Davey, J., Cell 1985.42: 457. 5 Taylor, I? M., Davey, J., Howland, K., Rothbard, J. B. and Askonas, B. A., Immunogenetics 1987. 26: 267. 6 Townsend, A. R. M., Rothbard, J., Gotch, F. M., Bahadur, G., Wraith, J. and McMichael, A . J., Cell 1986. 44: 959.

Antigen-specific activation of memory CD8+ T cells

1601

7 Pala, F! and Askonas, B. A., Immunogenetics 1986. 23: 379. 8 Engers, H. D. and MacDonald, H. R., Contemp. Top. Immunobiol. 1976. 5: 145. 9 Effros, R. B., Bennink, J. and Doherty, l? C., Cell. Immunol. 1978. 36: 345. 10 Cerottini, J.-C. and MacDonald, H. R., Annu. Rev. Immunol. 1989. 7: 77. 11 Vitetta, E. S., Berton, M.T., Burger, C., Kepron, M., Lee,W.T. and Yin, X.-M., Annu. Rev. Immunol. 1991. 9: 193. 12 Weiss, A , , in Paul, W. E. (Ed.), Fundamental Immunology, Raven Press, New York 1989, p. 359. 13 Bodmer, H. C., Pemberton, R. M., Rothbard, J. B. and Askonas, B. A , , Cell 1988. 52: 253. 14 Yap, K. L. and Ada, G . L., Immunology 1977. 32: 151. 1.5 Rotzschke, O., Falk, K., Deres, K., Schild, H., Norda, M., Metzger, J., Jung, G. and Rammensee, H.-G., Nature 1990. 348: 252. 16 Askonas, B. A., Mdllbacher, A. and Ashman, R. B., Immunology 1982. 45: 79. 17 Pilarksi, L. M., J. Exp. Med. 1977. 145: 709. 18 Ashman, R. B. and Miillbacher, A., J. Exp. Med. 1979. 150: 1277. 19 Widmer, M. B. and Bach, F. H., Nature 1981. 294: 750. 20 Von Boehmer, H., Kisielow, F!, Leiserson, W. and Haas, W.. J. Immunol. 1984. 133: 59. 21 Ahmed, R.. Butler, L. D. and Bhatti, L., J. Virol. 1986. 62: 2102. 22 Buller, R. M., Holmes, K. L., Hugin, A , , T. N. F., and Morse 111, H. C., Nature 1987. 328: 77. 23 Liu,Y. and Miillbacher, A ., Immunol. Cell Biol. 1989.67: 413. 24 Liu,Y. and Mullbacher, A ., Proc. Natl. Acad. Sci. USA 1990. 86: 4629. 25 Heeg, K., Steeg, C., Hardt, C. and Wagner, H., Eur. J. Immunol. 1987. 17: 229. 26 Kung, J.T., Castillo, M., Heard, F!, Kerbacher, K. and Thomas 111, C. A ., J. Immunol. 1991. 146: 1783. 27 Akbar, A. N., Salmon, M. and Janossy, G., Immunol. Today 1991. 12: 184. 28 Vitiello, A , , Heath, W. R. and Sherman, L. A , , J. Immunol. 1989. 143: 1512. 29 Sprent, J. and Schaefer, M., Immunol. Today 1989. 10: 17. 30 Sprent, J. and Schaefer, M., Immunol. Rev. 1990. 117: 213. 31 Andrus, L., Granelli-Piperno, A. and Reich, E., J. Exp. Med. 1984. 159: 647. 32 Erard, F., Corthesy, l?, Nabholz, M., Lowenthal, J.W., Zaech, F!, Plaetnick, G. and MacDonald, H. R., J. Immunol. 1985. 134: 1644. 33 Simon, M. M., Landolfo, S., Diamanstein, T. and Hochgeschwender, U., Curr. Top. Microbiol. Immunol. 1986.126: 173. 34 Horak, I. D., Gress, R. E., Lucas, l? J., Horak, E. M., Waldmann,T. A. and Bolen, J. B., Proc. Natl. Acad. Sci. USA 1991. 88: 1996. 35 Bretscher, I? and Cohn, M., Science 1970. 169: 1042. 36 Langman, R. E., The immune system. Evolutionary principles guide our understanding of this complex biological defense system, Academic Press, Inc., San Diego 1989, p. 209. 37 Liu,Y. and Miillbacher, A., Scand. J. Immunol. 1989.30: 277. 38 Perkins, D. L., Lai, M.-Z., Smith, J. A. and Gefter, M. L., J. Exp. Med. 1989. 170: 279.