Kinetics of H L A class I- and class 11-restricted antigen presentation

Eur. J. Immunol. 1992. 22: 2339-2345

Karel C. KuijpersO, Folkert J. van Kernenaden, Berend Hooibrinko, Jacques J. NeefjesO, Cees J. LucasoA, Rent! A. W. van Lieroe and Frank Miedemaoe Department of Clinical Viro-Immunology and Laboratory for Clinical and Experimental Immunology of the University of Amsterdam incorporated in the Central Laboratory of the Netherlands Red Cross Blood Transfusion Servicen, Amsterdam and Department of Cellular Biochemistry, The Netherlands Cancer Institutec, Amsterdam

HLA class I and I1 molecules present influenza virus antigens with different kinetics* Human leukocyte antigen (HLA) class I and class I1 molecules differ with respect to their intracellular pathways and the compartments where they associate with processed antigen. To study possible consequences of these differences for the kinetics of antigen presentation by HLA class I and class I1 molecules, we analyzed changes in the concentrations of free intracellular calcium ions in influenza virus-specific T cell clones after recognition of specific antigen/HLA complexes. HLA class 11-restricted viral antigen presentation by Epstein-Barr virus-transformed B lymphoblastoid cell lines (B-LCL) t o CD4+ T cell clones started within 1h and showed little variability, irrespective of antigen specificity or restriction element tested. In contrast, kinetics of viral antigen presentation by HLA class I molecules to CD8+ T cell clones were slower and differed for three antigen/HLA class I complexes tested. While B-LCL presented antigen by HLA-A2 and by HLA-B37 after at least 2 h, they only started to present antigen in the context of HLA-B7 after more than 4 h. This difference in kinetics did not correlate with differences in bulk transport rates of HLA-A2, HLA-B37, and HLA-B7, but seemed greatly influenced by differential rates of peptide generation. Brefeldin A treatment of B-LCL showed for both HLA class I and class I1 that de now synthesized HLA molecules were involved in antigen presentation. Thus, differences between intracellular pathways of HLA class I and class I1 molecules may result in different kinetics of antigen presentation.

1 Introduction Most antigens are only recognized by T cells after intracellular degradation to peptide fragments which are bound by major histocompatibility complex (MHC) molecules to be presented at the cell surface [1].These processes cause a certain lag time between uptake of antigen by, or virus infection of, antigen-presenting cells (APC) and its presentation to Tcells. Whereas MHC class I molecules predominantly present fragments of proteins generated intracellularly, e.g. de novo synthesized during a viral infection [2-41, class I1 molecules present peptides derived from proteins taken up by the endocytic route [2, 5-71. However, this division has become less strict, since several exceptions have been found for both classes of MHC molecules [8-111. MHC class I and class I1 molecules follow different pathways during the process of antigen presentation [12, 131. Class I molecules presumably bind peptide already present in the endoplasmic reticulum (ER) and follow the constitutive pathway to the cell surface. In contrast, class11 molecules are directed in the trans-Golgi reticulum to the

TI 99611

* This work was supported by a grant from the Dutch Society for A

2339

support of research on Multiple Sclerosis. Present address: Netherlands Organisation for Applied Scientific Research, Medical Biological Laboratory, Rijswijk. Senior fellows of the Royal Netherlands Academy for Arts and Sciences.

Correspondence: Karcl C. Kuijpers, Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, PO. Box 9190, NL-1006 AD Amsterdam, The Netherlands Abbreviations: ER: Endoplasmic reticulum din A

BFA: Brefel-

0 VCH Vcrlagsgesellschaft mbH, D-6940 Weinheim, 1992

endocytic compartment, where they lose the associated invariant chain (Ii), presumably after proteolysis of Ii [14-171, and bind peptides derived from endocytozed and degraded proteins. Loaded with peptide they continue on their way to the cell surface. The site of antigen entry, localization and intracellular concentration of peptide fragments may influence which class of MHC molecule will present it [ 181. In this study we have investigated whether kinetics of presentation of influenza virus antigens differ for HLA class I and class I1 molecules and whether differences exist for various influenza virus antigens or HLA-restriction elements. We have studied antigen presentation by measuring changes in concentration of free intracellular Ca2+ ([Ca2+],) in indo-1-loaded Tcell clones [19] specific for influenza virus. After conjugation of these Tcells with influenza virus-infected cells of EBV-transformed B lymphoblastoid cell lines (B-LCL), binding of the Tcell receptor (TcR) to its specific antigen/HLA complex leads within minutes to a series of activation events, one of which is a rise in [Ca2+]i[20,21]. Kinetics of antigen presentation by class I1 molecules are rapid and do not differ between different class I1 alleles, irrespective of the antigen presented, whereas antigen presentation by class I molecules occurs at a slower rate and depends on the class I allele and/or antigen involved.

2 Materials and methods 2.1 Cells For three donors, K46, K61, and K68, CD4+ and CD8+ T cell clones specific for influenza virus were generated from peripheral blood mononuclear cells, as described elsewhere [22] (see Table 1 for Tcell specificities). These 0014-2980/92/0909-2339$3.so+ .2s/o

2340

Eur. J. Immunol. 1992. 22; 2339-2345

K. C. Kuijpers, F. J. van Kemenade, B. Hooibrink et al

Influenza virus A/HK/S/68 was grown in 11-day-old chicken embryos at the National Influenza Center, incorporated in the Dept. of Virology, Erasmus University, Rotterdam (Prof. Dr. Masurel, Dr. Sprenger). Allantoic fluid of inoculated eggs was harvested and spun to remove cell debris. Virus titers were determined in a standard hemagglutination inhibition assay. Virus stocks were stored at - 70°C until use.

A/HK/S/68 (100 HAU/3 X 106 cells unless otherwise indicated), and cultured for various time periods at 37°C and 5% C02. To end further processing at the appropriate time point, cells were washed twice with ice-cold PBS, and thereafter they were resuspended in Hepes buffer (4 “C).To determine the influence of de novo synthesis of viral proteins on antigen presentation, influenza virus was UV irradiated, as described previously [S]. To inhibit egress from the ER of de novo synthesized HLA molecules, stimulator cells were pretreated with brefeldin A (BFA; Epicenter, Madison, MI; at concentrations indicated in Fig. 5 ) for 30 min, and subsequently infected in the presence of BFA and cultured for 4 h.Viability of BFA-treated B-LCL was always > %YO, as determined by trypan blue exclusion. As control, K46 and K61 B-LCL were sensitized by incubation with hemagglutinin peptide 236-252 (5 pg/ml, kindly provided by Dr. D. S. Burt, NIMR, London) instead of virus.

2.3 Media

2.5 Measurements of changes in [Ca2+]i

B-LCL were cultured and infected in RPMI 1640 supplemented with penicillin, streptomycin, 10% FCS, and 1% non-essential amino acids (all from Gibco Laboratories, Grand Island, NY). Measurements of [Ca2+]i were performed in Hepes buffer (132 mM NaCl, 6 mM KC1, 1 mM CaC12,l mM Na2HP04,1 mM MgS04,20 mM Hepes, 5 mM glucose, and 0.5% vol/vol human serum albumin).

T cell clones (up to 1 x 107/ml)were loaded for 20 min with 2 YM indo-1-AM (Molecular Probes, Eugene, OR) at 37°C in Hepes buffer pH 7.0. After adjusting the pH to 7.3 by addition of 9 vol of Hepes buffer (pH 7.3), cells were incubated for another 40 min, washed, resuspended to a concentration of lo6 celldm1 in Hepes buffer (pH 7.3), and kept at room temperature until use. To test the ability of infected or non-infected B-LCL to increase [Ca2+], of clonal Tcells, 1 x lo5 Tcells were pelleted together with 2 x lo5 APC and incubated at 37°C for 2 min. After resuspension of the pellet, fluorescence of indo-1-loaded Tcells was determined on a FACStar-Plus (Becton Dickinson, San Jose, CA) equipped with a sample delivery system

donors were HLA-typed. Donor K46 was typed A2,-; B7,14; Cw13,-; DR1,2; DQw1,-, donor K61 A24,32; B37,w61; DR5,-; DQw3,-, and donor K68 A26,28; B35,wS6; Cwl,w4; DR1,-; DQw1,-. Donor S52 was typed A1,2; B7,37; DR2,wlO; DQw3,-. B-LCL were prepared at our institute. 2.2 Influenza virus

2.4 Preparation of stimulator cells Autologous, HLA-matched, and mismatched B-LCL were used as APC. They were infected with influenza virus

Table 1. Antigen spccificity and HLA restriction of influenza virus-specific T cell clones

Donor

Clone

Phenotype

Antigen specificity

HLA restriction

K46

21 2c11 3A8 4H7 6D5 6H7 7Cl 8C11 5B3

CD4 cD4 CB4 CD4 CD4 CD4 CD4 cD4 CD8

ncuraminidase neuraminidase hemagglutinin neuraminidasc ncuraminidase neuraminidase hemagglutinin (236-252)”) neuraminidase non-structural 2 matrix protein (56-6t3)a)

CD4 CD4 CD4 CD4 CD4 CD8 CD8 CD4 CD4 CD4 CD4 cD4 CD4 CD4

matrix protein hemagglutinin (100-1 IS)*) neuraminidase tiemagglutinin (81-W)a) hemagglutinin (236-252)a) nucleoprotein (335-349)a) nuclcoproteio (335-349)”)

HLA-DRl HLA-DR2 HLA-DR2 HLA-DRl HLA-DRt W.A-DRl HLA-DQwl HLA-DR 1 HLA-B7 HLA-A2 HLA-DRS HLA-DR5 HLA-DR5 HLA-DQw3 HLA-DK5 HLA-B37 HLA-B37 HLA-DRl H’LA-DK1 HLA-DR1 HLA-DRl HLA-DKl HLA-DKl HLA-DRl

K6 1

22 1 5 7

K68

14 20 4A 1 CDWl 2 h

13 16 18 23 36

cns

matrix protein nucleoprotein matrix protein nucleoprotein hemagglutinin ncuraminidasc nucleoprotein

a) Antigenic epitope as determined with synthetic peptides.

Eur. J. Tmrnunol. 1992. 22: 2339-2345

Kinetics of HLA class I- and class II-restricted antigen presentation

in which cells were kept at 37 "C and magnetically stirred. Forward and sideward scatter were used to gate T cell/APC conjugates.The percentage of responder cells in conjugates and the 405/485-nm absorbance ratio of indo-1 fluorescence of each individual cell, taken as a measure for Tcell responsiveness 1231, were calculated using the kinetic software INCA [24]. Before each experiment, the FACStarPlus was calibrated using indo-l-coated latex beads (Flow Cytometry Standards Corp., NC). All Tcell clones tested had a basal 405/485-nm ratio of approximately 1, which indicated that they were not preactivated. Results were expressed as percentages of cells responding with a ratio above background level in conjugates between T cells and APC. In pre-run experiments, background levels of fluorescence were determined for T cells in the absence of APC and set at a level that included 95% or more of the cells. 2.6 Pulse-chase experiments Pulse-chase experiments were performed as described previously (251. Briefly, 2 x lo7 cells were cultured for 45 min in methionine-free RPMI 1640 medium and then pulsed with 500 pCi (= 18.5 mBq) [3sS]methionine for 10 min. Further incorporation of the label was stopped by addition of non-radioactive methionine to a final concentration of 1 mM. A t different chase times a constant volume was taken from the cell suspension to which lysis buffer containing Nonidet-P40 was added. HLA class I molecules were immunoprecipitated with monoclonal antibody W6/32 [26]. Immunoprecipitations were loaded on an isoelectric focusing (IEF) gel. One-dimensional IEF and autoradiography were performed as described [26].

3 Results 3.1 Analysis of antigen presentation by measurements of changes in [Ca2+Iiin influenza virus-specific T cell clones We first determined the amount of influenza virus necessary to obtain optimal T cell responses. Since T cell/APC 6H7

KG8-36

2341

conjugate formation is a prerequisite for this type of T cell activation, only the response of conjugated Tcells was analyzed and expressed as the percentage of responding Tcells in these conjugates. As a measure for responsiveness, the 405/485-nm ratio of indo-1 fluorescence of conjugated cells was calculated. Both percentages and ratio of indo-1 fluorescence changed within 2 min after conjugate formation and subsequent T cell stimulation through the CD3mcR complex (data not shown). Because increases in ratio parallelled those in the percentage of responder cells, we chose to present the results only as percentages of responder cells in T cell/APC conjugates. Invariably, T cell responses strictly depended on the presence of the correct HLA-restriction element and viral antigen. Furthermore, neither the numbers of conjugates betweenTcells and APC nor expression of adhesion molecules involved in conjugate formation (LFA-1, ICAM-1, and LFA-3) varied with the length of infection, UV irradiation of the virus, or BFA pretreatment of B-LCL (data not shown). It was not necessary to fix APC before use, because in this assay system, in contrast to proliferation or cytolytic assays, the time needed to measure T cell activation was negligible compared to the time required for antigen presentation. Unwanted side-effects of fixation could thus be avoided. As shown in Fig. 1, B-LCL infected with 100 HAU influenza virus induced highest responses in CD4+ (clones 6H7 and K68-36) and CD8+ (4A1) T cell clones (specificities of the clones used in this study are listed in Table 1). B-LCL infected with 25 HAU gave marginally lower responses, whereas those infected with only 6 HAU of influenza virus clearly were suboptimal stimulator cells for most clones. Since infection with more than 100 HAU of influenza virus did not accelerate or increase the CD4+ or CD8+ Tcell response significantly (data not shown), further infections were done with this amount of virus. Previous studies showed that UV irradiation of influenza virus decreased antigen recognition by CD8+ but not by CD4+ T cells [2, 81. We confirmed these differentially inhibitory effects at the level of changes in [Ca2+],,a very early indicator of Tcell activation. As shown in Table 2, B-LCL infected with UV-irradiated virus induced lower percentages (and lower ratios of indo-1 fluorescence, data not shown) of responder cells of CD8+ Tcell clones 5B3,

4A1

Figure 1. Antigen presentation by HLA class I or class I1 molecules is not accelerated by infection with increased amounts of virus. Autologous B-LCL were infected with 6 HAU (A), 25 HAU (O), or 100 HAU (0)of influenza virus for various periods of time. At the time pointsindicated they were testedfor the ability to increase the percentage of respondingTcells inTcelllAPCconjugates. Specificitics of the CD4+ (6H7, K68-36) and CD8+ (4A1) Tcell clones are listed in Table 1.

Eur. J. Immunol. 1992. 22: 2339-2345

K. C. Kuijpcrs, F. J. van Kemenade. B. Hooibrink et a1

2342

B

A

C

Figure. 2. Kinetics of presentation of influenza virus antigen by HLA class I1 molecules is rapid, irrespective of the class I1 allele or viral protein tested. Antigen presentation by autologous B-LCL to panels of CD4+ Tcell clones (seeTable 1for specificities) was expressed as ( 0 ) 2 1 , (0)6H7, (A) 7C1, ( A ) 2 C l l , (0) 3A8, (m)4H7, percentage rcsponderTcellsinTcell/APCconjugates.TcellclonesusedareinA: (V)6D5; in B: (0)K68-2. ( 0 )K68-6, (A) K68-13, (A)K68-18, (0)K68-23; in C: (0)K61-1, ( 0 )K61-5, (A) K61-7, (A)K61-14, (0) K61-20. In further experiments clones K61-5 and K61-14 reached percentages of respondingTcells comparable to those of clones K61-1, K61-7 and K61-20 (data not shown).

K46-22, and 4A1 than B-LCL infected with untreated virus. In contrast, both types of infected APC stimulated CD4+ T cell clones equally well. Thus also in this readout system, UV irradiation of the virus inoculum diminished Presentation of viral antigens by HLA class I but not by class I1 molecules. 3.2 Antigen presentation by HLA class I1 molecules Kinetics of antigen presentation by HLA class I1 molecules was studied in panels of CD4+ T cell clones with different antigen specificities and restriction elements (Table 1). As shown in Fig. 2, when B-LCL were infected for 1 h, they already induced a response of all CD4+ T cell clones. After 2 h of infection, B-LCL uniformly elicited approximately half-maximal responses. When infected for 4 h or longer, they triggered maximal percentages of respondingT cells in all clones. HLA-DR1 on K46 B-LCL (Fig. 2 A ) presented neuraminidase to clones 21, 4H7, 6D5, and 6H7 at comparable rates, which resembled that of HLA-DR2 presenting neuraminidase to clone 2Cl1. HLA-DR1 on homozygous K68 B-LCL (Fig. 2B) showed also similar

kinetics of presentation of various antigens (matrix protein, nucleoprotein, hemagglutinin, and neuraminidase) to the Tcell clones tested (K68-2, K68-13, K68-6, K68-18, and K68-23, respectively). Furthermore, HLA-DRS on K61 B-LCL (Fig. 2 C) presented matrix protein, neuraminidase, and hemagglutinin at comparable rates to clones K61-1, K61-7, and K61-20, respectively. Comparable results were obtained for presentation by HLA-DQwl and -DQw3. Thus, kinetics of antigen presentation by HLA class I1 molecules showed surprisingly little variability, irrespective of the influenza virus antigens or restriction elements tested. 3.3 Antigen presentation by HLA class I molecules Presentation of antigen by HLA class I molecules occurred later than that by class I1 molecules. As shown in Fig. 3, B-LCL started to present nucleoprotein by HLA-B37 to Tcell clones 4A1 and K61-CD8/1 after more than 2 h of infection, whereas they induced a strong response in HLA-DR1-restricted nucleoprotein-specific T cell clone K68-36 already within 1 h. Likewise, matrix protein was

Table 2. Effect of UV irradiation of influenza virus on antigen presentation by HLA class I and I1 molecules

Clone

Phcnotypc Tinil: (h)

SB3*I

CD8

K46-22"'

CIM

K 4 6 B-LCL") K61 B-LCL') AHK AHK-UV AHK rWK-UV

4 6 8 1

2 4 6

7C 1h, 1A I

"'

K6 1 -20b)

CD4 CD8

CD4

4 2 4

-

6

11 23 25

2

88

4 6

81 89

a) K46 and KDhl B-LCL were infected with 100 HAU UV-irradiated or untreated influenza virus AIHKI8/68 for the time periods indicated. b) See Table 1 for clonal specificities. c) Results are expressed as percentage responder T cells in conjugates. d) Not done.

Kinetics of HLA class I- and class II-restricted antigen presentation

Eur. J. Immunol. 1992. 22: 2339-2345

2343

slowly to infected B-LCL than CD4+ Tcell clones did, they finally reached percentages of responder cells and ratios of indo-1 fluorescence comparable to CD4+ clones (Table 2 and data not shown). Previously, differences in association of HLA class I heavy chains with fiz-microglobulinand subsequent appearance at the cell surface have been described [25]. To determine whether the bulk transport rates of the class I restriction elements analyzed in this study correlated with the observed differences in kinetics of antigen presentation, the appearance of HLA-A2, HLA-B7, and HLA-B37 in the trans-Golgi was analyzed, which is correlated with cell d 5 1 4 6 8 surface expression [12, 251. Addition of sialic acids is time (hours) considered to occur in the trans-Golgi and results in a shift to a more acidic position of class I molecules on oneFigure 3. Kinetics of antigen presentation by HLA class I moledimensional IEF. In B-LCL expressing HLA-B7 (K46), cules occurs at a slower rate than that by class11 and differs HLA-B37 (K61), or both (S52), sialylation of HLA-B7 was between class I alleles. Autologous B-LCL presented nucleoprotein by HLA-B37 to CD8+ Tcell clones 4 A 1 ( 0 ) and K61-CD8/1 completed after 35 min, that of HLA-B37 took 2 h, while (a),and by HLA-DR1 to CD4+ clone K68-36 (0),while they sialylation of HLA-A2 was even slower, as shown in Fig. 4. presented matrix protein by HLA-A2 to CD8+ Tcell clone K46-22 Bulk transport rates were not influenced by influenza virus (A),and by HLA-DRl to CD4+ clone K68-13 (A). Non-structural infection ([27], and data not shown). Thus, bulk transport protein-2 was presented by HLA-B7 to CD8+ Tcell clone 5B3 (0). rates of HLA-A2, HLA-B7, and HLA-B37 could not be Results are expressed as percentage responder Tcells inT cell/APC correlated with kinetics of antigen presentation by these conjugates. restriction elements. )

presented sooner after the start of infection by HLA-DR1 than by HLA-A2. Furthermore, HLA-A2, HLA-B37, and HLA-B7, the class I restriction elements studied, differed in kinetics of antigen presentation.Whereas B-LCL started to present matrix protein by HLA-A2 to clone K46-22 or nucleoprotein by HLA-B37 to clones 4A1 and K61-CD8/1 after approximately 2 h, they only began to induce comparable responses in clone 5B3, restricted by HLA-B7 and probably specific for non-structural protein-2, after 4-6 h of infection (Fig. 3, and Table 2).This difference in kinetics was consistently found, also when e.g. S52 B-LCL, expressing HLA-A2, HLA-B37, and HLA-B7,were used (datanot shown). Although CD8+ Tcell clones responded more

3.4 Both HLA class I and class I1 molecules require de novo synthesized HLA molecules for antigen presentation

Presentation of antigen by HLA class I molecules was slower than that by class I1 molecules. This difference in kinetics could be related to a differential use of de novo synthesized vs. preformed HLA molecules. To determine whether viral antigens were presented by de novo synthesized HLA molecules, as recently demonstrated for class I as well as class I1 molecules [28-311, or by (recycling) molecules from a pre-existing pool, as has been suggested

B

A

C

W6l32 time

0

10

20

30 60

180

0

10 20

30

00

180

0

10 20

30

60

180

0 37

1:E Figure 4. A biochemical analysis of the rate of intracellular transport of class I molecules. Cells (K46, K61, or S52 B-LCL) were labeled with ["S] methionine and chased for the periods of time indicated above the figure. HLA class I molecules were immunoprecipitated with monoclonal antibody W6/32. In a one-dimensional IEFgel, the positions of HLA-A2, -B7, and -B37, as well as their sialic acid-containing forms are indicated for K46 (A), S52 (B), and K61 (C).The rate of sialylation of HLA-B7 is faster than that of HLA-B37, which is faster than that of HLA-A2. Anode is at the bottom.

2344

K. C. Kuijpers, F. J. van Kemenade, B. Hooibrink et al.

for class I1 molecules 1321, we incubated B-LCL with BFA, a compound shown to inhibit protein transport from the E R to the Golgi complex [33, 341. When BFA was present during infection, it inhibited antigen presentation to both CD4+ and CD8+ clones in a dose-dependent way (Fig. 5). This inhibition was observed for all viral proteins (matrix protein, nucleoprotein, hemagglutinin, and neuraminidase) and restriction elements tested (HLA-A2, -B37, -DR1, and -DRS). Furthermore, BFA inhibited antigen presentation of UV-irradiated as well as untreated influenza virus equally well (data not shown). BFA did not affect recognition of peptide-sensitized B-LCL (Fig. 5 ) , which is assumed to depend on binding or exchange of peptides on preformed HLA molecules at the cell surface.These results showed that only de novo synthesized HLA class1 and class IT molecules are involved in antigen presentation.

v

c,

06

1

1 L I-)

75

L i , e i i l i 1 1 1A ( pM

Figure 5. BFA inhibits HLA class I- and class 11-restricted antigen presentation without affecting peptide presentation. A: BFA inhibits antigen presentation by B-LCL infected with influenza virus for 4 h to CDX+ T cell clones 4A1(0) and K46-22 (V),and to CD4+ clones K61-7 ( O ) , K61-20 (A), K6X-2 (A),K68-6 (m), and K6X-16 (0).The response of CD4+ Tcell clone K61-20 (V) to autologous B-LCL incubated with hemagglutinin peptide 236-252 was not inhibited by BFA (see Table 1 for specificities).

4 Discussion In this study, we have determined whether cell biological differences between HLA class I and class I1 molecules have consequences for the kinetics of presentation of influenza virus antigens and whether differences in kinetics exist for various viral antigens or HLA restriction elements. HLA class TI molecules present viral antigens rapidly. The lag time between infection and antigen presentation is approximately 1 h for various class I1 alleles, and is not shortened when more than 100 HAU influenza virus is used for infection. Previously, Roosnek et al. [19] have found a lag time of 1 h for the presentation of tetanus toxoid internalized via specific surface immunoglobulin. Thus, kinetics of class 11-restricted antigen presentation by BLCL occurs at comparable rates after viral infection or after antigen uptake via specific surface immunoglobulin. All HLA class I1 alleles tesed in this study (HLA-DR1, -DR2, -DRS, -DQwl, and -DQw3), show comparable

Eur. J. Immunol. 1992. 22: 2339-2345

kinetics of presentation for various influenza virus antigens (hemagglutinin, neuraminidase, matrix protein, and nucleoprotein). Although class I1 molecules can also bind peptides during their biosynthesis [lo, 35,361, they usually associate in the endocytic compartment with peptides derived from exogenous proteins internalized by endocytosis [l]. Usage of UV-irradiated influenza virus clearly demonstrates that rapid antigen presentation by class I1 molecules only requires endocytosis of the virus but no de novo synthesis of viral proteins, as described previously [2]. HLA class I1 molecules associated with the invariant chain (Ii) [13, 37-39] may reach the endocytic compartment already within 1 h after biosynthesis 112, 401. Class I1 molecules are directed to this compartment by a sorting signal in the cytoplasmic tail of associated Ii 141, 421. After Ii has been proteolyzed, e.g. by cathepsin B 116, 171, peptides derived by processing of endocytosed antigens can bind in the groove. The capacity of class I1 molecules associated with Ii to bind peptides is very low, but increases dramatically after dissociation of Ii [15, 16,431. Possibly, Ii bound to class I1 not only determines the direction but also the rate of transport of class I1 molecules, and proteolysis of Ii may be the rate-limiting step. Apparently, major differences in peptide generation and transport of antigen/class I1 complexes from the endocytic compartment to the cell surface do not exist in B-LCL. Our results suggest a uniform and very efficient binding of antigen by class I1 molecules in the endocytic compartment. BFA not only inhibits class I- but also class IIrestricted antigen presentation, as reported previously [28-311. These findings indicate that de novo synthesized class I1 molecules are involved in the presentation of various influenza proteins, extending earlier results on matrix protein [30]. Furthermore, BFA also inhibits class 11-restrictedpresentation of UV-irradiated virus, confirming inhibitory effects of BFA on presentation of exogenous antigens such as hen egg-white lyzozyme, chicken ovalbumin, or pigeon cytochrome C [31]. Kinetics of antigen presentation by HLA class I molecules is slower than that by class I1 molecules, and seems more heterogeneous. These differences in kinetics between HLA class I and class I1 molecules may be related to the different localizations where they associate with peptides. Furthermore, these differences may be influenced by the time required for infection and for viral protein synthesis and subsequent peptide generation. For presentation by class I molecules, proteins or their degradation products have to accumulate in the cytosol [8, 91. In the case of influenza virus infections, de novo synthesis of viral proteins is required [2], which only occurs after cytosolic insertion of viral RNA, following fusion between viral and endosomal membranes. As shown here and also described previously 12, 81, UV irradiation of virus inhibits class I-restricted antigen presentation. Most likely, peptides derived by degradation of de novo synthesized proteins associate with class I molecules early during class I synthesis in the ER, although this has not yet been proven unambiguously.This association allows correct assembly of class I molecules and further transport to the cell surface 144, 451. Many of those peptides are derived from cytosolic proteins and need to be transported over the E R membrane by a putative peptide pump 1461. Although the number of class I restriction elements tested is limited, the results suggest a heterogen-

Eur. J. Immunol. 1992. 22: 2339-2345

Kinetics of HLA class I- and class 11-restricted antigen presentation

eity in kinetics of class I-restricted antigen presentation. This heterogeneity may depend on efficiencyof transport of antigen/HLA complexes to the cell surface, subcellular localization of antigen, rates of antigen synthesis and proteolytic degradation, or otherwise on differences in TcR affinity. Although differences exist for HLA-A2, -B7, and -B37 in the association of HLA class I heavy chain with (32-microglobulin and subsequent cell surface expression [25],we cannot correlate these differences in bulk transport rates with kinetics of antigen presentation by these restriction elements. On the other hand, most of influenza virus proteins are synthesized early after infection [47], except non-structural protein-2 which is found only in infected cells several hours after infection [48]. Presentation of a peptide derived from non-structural protein-2 may explain the considerably slower kinetics of HLA-B7-restricted presentation observed in this study. Alternatively, differences in location and rate of peptide generation may exist. Thus, peptides may be rate limiting in the transport of specific antigen/HLA class I complexes to the cell surface. Our results indicate that differences in peptide acquisition and intracellular sorting pathways lead to kinetic differences between HLA class I and class I1 molecules with respect to their ability to present foreign antigens toTcells. Extrapolated to the in vivo situation, during secondary responses, infected APC first initiate CD4+ helper T cell responses before being eliminated by CD8+ cytotoxic Tcells. Temporal dissociation of activation of CD4+ and CD8+ Tcells may influence an efficient timing of both humoral and cellular immune responses. We would like to thank Ms. M. Stuiver of the Virology Dept. of the Erasmus University for growing of the virus, Dr. D. S. Burt for providing the hernagglutinin peptide, and Ms. G. Damhuis for secretarial assistance. Received September 29, 1991; in final revised form May 26, 1992.

5 References 1 Townsend, A. and Bodmer, H., Annu. Rev. Immunol. 1989. 7: 601. 2 Morrison, L. A., Lukacher, A. E., Braciale,V. L., Fan, D. P. and Braciale, T. J., J. Exp. Med. 1986. 163: 903. 3 Townsend, A. R. M., Gotch, F. M. and Davey, J., Cell 1985.42: 457. 4 Townsend, A. R. M., Bastin, J., Gould, K. and Brownlee, G. G., Nature 1986. 234: 575. 5 Ziegler, K. and Unanue, E. R., J. Irnmunol. 1981. 127: 1869. 6 Ziegler, H. K. and Unanue, E. R., Proc. Natl. Acad. Sci. U S A 1982. 79: 175. 7 Shimonkevitch, R., Kappler, J., Marrack, P. and Grey, H., J. Exp. Med. 1983. 158: 303. 8 Yewdell,J.W., Bennink, J. R. andHosaka,Y., Science 1988.239: 637. 9 Moore, M. W., Carbone, F. R . and Bevan, M. J., Cell 1988.54: 777. 10 Jaraquemada, D., Marti, M. and Long, E. O., J. Exp. Med. 1990. 172: 947. 11 Rudensky, A.Y., Preston-Hurlburt, P., Hong, S.-C., Barlow, A. and Janeway Jr., C. A., Nature 1991. 353: 622.

2345

12 Neefjes, J. J., Stollorz,V., Peters, P. J., Geuze, H. J. and Ploegh, H. L., Cell 1990. 61: 171. 13 Peters, P. J., Neefjes, J. J., Oorschot,V., Ploegh, H. L. and Geuze, H. J., Nature 1991. 349: 669. 14 Blum, J. S. and Cresswell, l?, Proc. Natl. Acad. Scz. U S A 1988. 85: 3975. 15 Roche, P. A. and Cresswell, P., Nature 1990. 345. 615. 16 Roche, P. A. and Cresswell, P., Proc. Natl. Acad. Scz. USA 1991. 88: 3150. 17 Reyes,V., Lu, S. andHumphreys, R. E., J. Imrnunol. 1991.146: 3877. 18 Morrison, L. A., Braciale,V. L. and Braciale,T. J., J. Immunol. 1988. 141: 363. 19 Roosnek, E., Demotz, S., Corradm, G. and Lanzavecchia, A , , J. lmrnunol. 1988. 140: 4079. 20 Tsien, R. Y., Pozzan, T. and Rink, T. J., Nature 1982. 295: 68. 21 Rabinovitch, P. S., June, C. H., Grossman, A. and Ledbetter, J. A , , J. Immunol. 1986. 137: 952. 22 Kuijpers, K. C., Treep-vanLeeuwen, P., Miedema, F. and Lucas, C. J., Eur. J. Immunol. 1991. 21: 1453. 23 Grynkiewics, G., Poenie, M. and Tsien, R., J. Biol. Chem. 1985. 260: 3440. 24 Keij, J. F., Griffioen, A. W., The, T. H. and Rijkers, G. T., Cytometry 1989. 10: 814. 25 Neefjes, J. J. and Ploegh, H., Eur. J. Immunol. 1988.18: 801. 26 Neefjes, J. J., Breur-Vriesendorp, B. S.,Van Seventer, G. A , , Ivanyi, P. and Ploegh, H. L., Hum. Immunol. 1986. 16: 169. 27 Rotteveel, F.T. M., Neefjes. J. J., Ploegh, H. L. and Lucas, C. J., Hum. Immunol. 1989. 26: 199. 28 Yewdell, J. W. and Bennink, J. R., Science 1989. 244: 1072. 29 Nuchtern, J. G., Bonifacino, J. S., Biddison, W. E. and Klausner, R. D., Nature 1989. 339: 223. 30 Nuchtern, J. G., Biddison, W. E. and Klausner, R. D., Nature 1990. 343: 74. 31 Adorini, L., Ullrich, S. J., Appella, E. and Fuchs, S., Nature 1990. 346: 63. 32 Reid, P. A. and Watts, C., Nature 1990. 346: 655. 33 Lippincott-Schwartz, J., Donaldson, J. G., Schweizer, A., Berger, E. G., Hauri, H. €!,Yuan, L. C. and Klausner, R. D., Cell 1990. 60: 821. 34 Serafini,T., Stenbeck, G., Brecht, A., Lottspeich, F., Orci, L., Rothman, J. E. and Wieland, F. T., Nature 1991. 349: 215. 35 Eisenliohr,L. C. and Hackett, C. J., J. Exp. Med. 1989.169: 921. 36 Jacobson, S., Sekaly, R. P., Jacobson, C. L., McFarland, H. F. and Long, E. O., J. Virol. 1989. 63: 1756. 37 Guagliardi, L. E., Koppelman, B., Blum, J. S., Marks, M. S., Cresswell, P. and Brodsky, F. M., Nature 1990. 343: 133. 38 Koch, N., Lipp, J., Pessara, U., Schenck, K., Wraight, C. and Dobberstein, B., Trends Biochem. Sci. 1989. 14: 383. 39 Cresswell, I?, Proc. Natl. Acad. Sci. U S A 1985. 82: 8188. 40 Machamer, C. E. and Cresswell, P., J. Immunol. 1982. 129: 2564. 41 Lotteau,V. ,Teyton, L., Peleraux, A., Nilsson.T., Karlsson, L., Schmid, S. L., Quaranta,V. and Peterson, P. A., Nature 1990. 348: 600. 42 Bakke, 0. and Dobberstein, B., Cell 1990. 63: 707. 43 Teyton, L., OSullivan, D., Dickson, P. W., Lotteau,V., Sette, A., Fink, P. and Peterson, P. A., Nature 1990. 348: 39. 44 Ljunggren, H.-G., Stam, N. J., Ohlen, C., Neefjes, J. J., Hoglund, P., Heemels, M.-T., Bastin, J., Schumacher,T. N. M., Townsend, A., Karre, K. and Ploegh, H. L., Nature 1990.346: 476. 45 Schumacher,T.N. M., Heemels, M., Neefjes, J. J., Kast,W. M., Melief, C. J. M. and Ploegh, H. L., Cell 1990. 62: 563. 46 Spies, T. and DeMars, R., Nature 1991. 351: 323. 47 Lamb, R. A. and Choppin. P.W., Annu. Rev. Biochem. 1983. 52: 467. 48 Lamb, R. A., Etkind, P. R. and Choppin, P. W., Virology 1978. 91: 60.