Immunology Today, voL 7, No. 9, 1986
-rostrum
Processing of viral antigens and presentation to class II-restricted T cells Some antigens require intracellular processing by antigen presenting cells before being presented to T cells in conjunction with surface major histocompatibility complex antigens. The whole mechanism of these processing events is not known and in this article, Kingston Mills puts forward arguments for alternative routes of antigen processing, with particular reference to recognition of viral proteins by class II-restricted T-cell clones.
Activation of antigen-specific helper T (Th) cells requires recognition of the foreign antigen in association with class II major histocompatibility complex (MHC) gene products on the surface of an antigen presenting cell (APC) 1'2. Certain foreign macromolecules may require internal handling or processing by the APC before presentation of the antigenic determinant to the T cell 3-5. If we define antigen processing as the cellular and molecular events which occur between antigen binding to APC and subsequent recognition by the T cell, then it is evident from the literature that the nature of these events is uncertain, in particular the extent of biochemical change in the antigen.
Lysosomal antigen processing Unlike B cells which recognize determinants on the tertiary structure of a protein, class II-restricted T cells are thought to see mainly linear determinants on the primary structure of peptides, protein fragments or denatured antigen 6~1°. There is indirect evidence that large globular antigens require processing steps which involve denaturation and proteolysis in the lysosomal compartment of the APC: (1) a lag period, where metabolic events occur, exists between antigen binding to the APC and recognition by the T cell3,S; (2) antigens are internalized and transported to compartments within the APC which are resistant to protease strippingS'l°; (3) agents which inhibit lysosomal function by raising the pH (chloroquine and ammonia) or inhibit proteolytic enzymes (leupeptin) prevent presentation of globular antigens but not fragments or peptides7'9,11; and (4) aldehyde fixation of APC inhibits presentation of globular antigen but not peptides 7'9'12'13. There is increasing experimental evidence which challenges this original concept of antigen processing. For example, it is now clear that a wide variety of cells other than phagocytic macrophages can present antigen: resting and immune B cells 1 3,1 4 , B-cell lines or tumours 5,1 3 - 15, dendritic cells 16,17 and la + T cells 19 have all been shown to function as APC. Many of these cells either are not phagocytic, have poor endocytotic abilities, or have few lysosomes. Membrane preparations or liposomes containing the appropriate MHC protein and antigen are also capable of activating class II-restricted T cells ls'19 Experiments with the lysosomotropic agents have shown that these drugs do not always inhibit presentation of 260
NationalInstitutefor MedicalResearch,LondonNW7 1AA, UK
KingstonH.G. Mills globular antigens 5,11'~5'17. Furthermore, chloroquine or ammonia can inhibit other cell functions such as membrane receptor recycling or biosynthesis of la molecules 2c~22. Studies with radiolabelled antigens 23 have also suggested that a proportion of native antigen is not internalized but remains on the surface of the APC, where it could then be recognized by T cells after association with la.
Recognition of native structure by influenza virus-specific T-cell clones We have recently provided evidence that certain class II-restricted T cells can recognize conformational features of a native antigen 24. T-cell recognition was studied using influenza virus-specific Th-cell clones established from individual mice primed 3-6 months earlier by infection with X31 influenza virus. From a panel of clones specific for either the haemagglutinin (HA) surface glycoprotein or the internal matrix protein (MP), seven distinct HA-specific clones failed to respond to an X31 mutant virus (R19) with a single amino acid substitution (His to Arg) at position 17 of HA1 (Fig. 1)2s. Each of the seven clones also failed to recognize pH5-treated virus, a synthetic peptide corresponding to residues 1-27 of HA1 or tryptic fragments consisting of HA1 residues 28-328 (tops) and the remainder of the virion including residues 1-27 of HA1 (aggregate). A further twenty HA-specific clones responded to R19 and to tops and showed a positive or slightly reduced response to pH5-treated virus, whereas two MP-specific clones recognized R19, aggregate and pH5-treated virus. It is unlikely that the R19 negative clones are specific for an epitope in the region of position 17 since aggregate and peptide 1-27 were not recognized. It could be argued that the His to Arg substitution in R19 interferes with the ability of the H-2 k APC to process the antigen, or of the class II molecules to associate with an altered la interaction site. However, since H-2 k APC could present R19 to other influenza virus-specific H-2 k class IIrestricted clones, this explanation seems unlikely. A study of the structure of the R19 mutant virus suggested that the single amino acid substitution at position 17 affects the conformational stability of the HA molecule 25. By testing the specificity of the clones for natural variant viruses of known amino acid sequence, the recognition site of five of the seven clones was tentatively mapped to a determinant in the interface of the HA trimer in the region of residues 213 and/or 208. The interface antibody combining sites are antigenically and structurally altered in tops and in pH5-treated virus as a result of an irreversible conformational change which affects the integrity of the globular head region of the HA trimer 26. Therefore it was concluded that at least a proportion of class II-restricted T cells recognize conformational determinants on the three dimensional structure of the native HA molecule. This suggests that an alternative to the lysosomal degradative route of the ~) 1986,ElsevierSciencePublishe~B.V.,Arnsterdam 0167- 4919/86/$02.00
Immunology Today, voL 7, No. 9, 1986
ros/rum
Interface Antibody Combining Site H A Trimer
HA1
R19 (X31 M u t a n t Virus)
-- HA2
N a t i v e Virus + Chlor o q u i n e / A m m o n i a
/Leupeptin
~Viral N e t i v e X31 Virus
p H 5 Virus
Aggregate
H A - 1 Tops
Conformational-dependent H A - s p e c i f i c Clones
"1"
Other
-~,
,I- / - -
--
~)-
"~"
"0"
4"
~1"
--
"1"
HA Specific Clones
Internal P r o t e i n specific Clones
.
.
.
.
"~"
4"
Fig.1. Recognitionof viralproteinsby influenzavirus-specificclassII-restrictedT-cellclones. Theproliferativeresponseof a panelof T-cellcloneswas testedwith an optimumconcentrationof nativevirus,pHS-treatedvirus, viraltryptic fragments(HAI tops and aggregate)and an X31 mutant virus (RIg). Cloneswere divided into three groups. internal MP-specific; conformationaldependantHA-specific; and other HA-specificcloneswith a rangeof spedfidtiesinduding thosewhich recognizesyntheticpeptidesof HA29.Alsoshownare the responsesto the nativevirususingAPC pre-incubatedfor 1 h with chloroquine(0. I mM), or pre-incubatedand pulsedwith antigenfor 2 h in the presenceof ammoniumchloride(1.0 mM) or leupeptin(1.0 mM).
antigen processing does exist for certain HA-specific T cells. The low pH of the endosome or lysosome alone would alter the antigenic determinants recognized by these T cells. Therefore a route which can circumvent irreversible pH5-induced changes in structure and lysosomal proteolysis must be envisaged.
Effect of APC fixation on antigen presentation In an attempt to define possible mechanisms of antigen processing for HA- or MP-specific Th-cell clones, we tested the effect of cell fixation on antigen presentation. We failed to demonstrate presentation of native virus or BHA to any of the T-cell clones tested using glutaraldehyde or paraformaldehyde fixed splenic APC (unpublished). Proliferation was induced in a T-cell line, but only when the APC were pulsed with antigen before fixing. Furthermore, T-cell clones specific for a peptide of HA1 (48~8) failed to respond to the peptide with fixed APC. It is unlikely that a peptide as short as 20 amino acids would require further degradation for recognition. The fixed cells were not toxic since they did not inhibit T-cell proliferation in the presence of non-fixed APC. It is possible that lack of presentation by fixed cells may reflect a requirement for a feeder effect or IL-1 release by metabolically active APC27 or it may be due to distortion of the T-cell receptor or antigen interaction site on the la molecule. The majority of previous studies that demonstrated recognition of peptide antigens by fixed APC have involved measurement of IL-2 release by T-cell hybridomas as an index of T-cell activation. The requirements for induction of proliferation by T-cell clones may be more stringent.
Effect of lysosomotropic agents Experiments with lysosomotropic agents did provide evidence for alternative routes of antigen processing (unpublished). Chloroquine, ammonium chloride and leupeptin all inhibited presentation of whole virus or of purified MP to MP-specific clones. However, the effects on HA-specific clones was variable. Recognition of peptide 48-68, purified HA or native virus by peptidespecific clones was not inhibited to any great extent and in some cases (responses to virus or HA in presence of
leupeptin) was augmented. Residues 48-68 are on an exposed antibody combining region of the HA molecule, therefore it is possible that the T-cell clones may see a determinant on the unprocessed native virus, the structure of which can be adopted by the free synthetic peptide. Augmentation of proliferation in the presence of proteolytic inhibitors has previously been suggested to result from preservation of antigenic determinants in the native structure which might be lost during proteolysis 28. In the case of the conformational-dependent clones there was slight (10-40%) inhibition of proliferation to whole virus in the presence of chloroquine and leupeptin. This may argue for a lysosomal route of processing; however chloroquine may also affect the alternative routes of antigen processing described below.
Possible mechanisms for processing of influenza virus antigens It is evident from the findings with influenza virusspecific Th clones that more than one route exists for processing of viral antigens. Four possible alternatives are schematically described in Fig. 2. The first step in each route involves binding of the viral HA to a receptor on the APC membrane. This receptor may be a sialic acid containing membrane receptor or a specific anti-HA antibody on the surface of a B cell. In the case of exogenous peptide antigen, binding may simply involve non-specific sticking to the APC membrane. Route A is the classical lysosomal route and involves endocytosis, fusion of the endosome with a lysosome, denaturation and proteolytic digestion in the acidic environment of the lysosome and excretion of fragments to the cell surface. This route would be blocked by lysosomotropic agents and may be the one used for processing of internal viral proteins. However, it is just as likely that an alternative internal pathway is employed (route B). In this case the contents of the endosome are dissociated by a mechanism which does not require entry into a lysosome. This mechanism may involve release of the viral contents into the cytoplasm. Non-lysosomal pathways of proteolysis, such as the ubiquitin-associated system, have been described for several proteins and may also be blocked by chloroquine or ammonium chloride through their effect on endocytosis2°
261
Immunology Today, vol. 7, No. 9, 1986
rostru I Antigen Binding to Specific Receptor
lI
111
Endocytosis
I~
Denaturation and Proteolysis
Exocytosis
~" Association with la
31I Ag-la-TcR Ternary Complex Formation
Exogenous Antigen Fragments Viral
T Cell
Lysosomal Degradation Endosome
T Dell
Lysosom~ NP
Native Virion
# Cytoplasmic
Fusion of Viral and
Native HA Molecule
'Membrane Processing'
(~
Exogenous BHA
Fig.2. Schematicrepresentationof alternativeroutesfor processingof influenzavirusantigen. Fourpossibleroutes(A-D)areshownwithsixindividualsteps(I-VI)for routeA, manyof whicharebypassedin the otherroutes.A: internallysosomalprocessing.B: non-lysosomalinternaldegradation.C: infectionof the APCby the virusfollowedbyexpressionof nativeglycoproteinson the cellsurfacemembrane.D: presentation of nativeantigenwithoutinternalization.
Route C requires infection of the APC by the virus. During infection fusion of viral and endosomal membranes occurs permitting release of the viral RNA for transcription and translation. Newly synthesized viral glycoproteins are excreted via the Golgi apparatus and expressed on the external surface of the cell membrane, where they could then be recognized by the T cell in their native conformation. The internal pathways are bypassed in route D. Here native virus or HA is presented to the T-cell without internal processing. Certain antigens may undergo proteolysis at the cell surface membrane 28 ; however this is unlikely to occur with the relatively proteolytic-resistant native influenza virion. Therefore the only step between virus binding to the cell surface receptor and T-cell recognition may involve association of the antigen with class II molecules. This association may be a loose interaction which is stabilized after binding of the T-cell receptor or it may only occur in the presence of the T-cell receptor.
262
Evidence for alternative routes of antigen processing We may conclude from the studies with influenza virus-specific Th-cell clones that internalization and lysosomal degradation of globular antigens are not essential features of antigen processing. Alternative mechanisms which permit recognition of native structures by T cells clearly exist. Interestingly, recognition of the conserved MP, which is enclosed within the viral membrane,
appears to require an intracellular proteolytic processing step, whereas the exposed HA molecule can be recognized in its native shape. Therefore the specificity of the T cell may determine the requirement for antigen processing. The previous assumption that class II-restricted T cells recognize linear determinants on processed antigen was based on studies with conserved proteins and with T cells derived from mice immunized with antigen in adjuvant. We have used T-cell clones established from mice primed by infection to demonstrate a previously unrecognized diversity in class II-restricted T-cell recognition of antigen, including specificity for determinants in exposed variable antibody combining regions of the HA molecule 29. Aggregation of an antigen with adjuvant, which is required for effective priming with soluble globular antigens, renders it less susceptible to recognition in its native conformation. Immunization by infection overcomes the requirement for priming in adjuvant and may explain previous failures to detect Th-cell recognition of tertiary structures. Furthermore memory T cells may employ a different mechanism for antigen processing than that required for a primary response. Antigenspecific B cells may focus antigen in vivo, via their immunoglobulin receptors, for recognition by antigenspecific T cells. Immune B cells have been shown to function as effective APC in vitro at lower antigen concentration than that required for macrophages 14'3°. A number of studies have shown that a functional
ImmunologyToday,vol. 7, No. 9, 1986
ros/rum
References 1 Kappler, J.W. and Marrack, P.C.(1976) Nature (London) 262, 797 2 Schwartz, R.H. (1985)Annu. Rev. Immunol. 3, 237 3 Ziegler, K. and Unanue, E.R.(1982) J. ImmunoL 127, 1869 4 Unanue, E.R.(1982)Annu. Rev. ImmunoL 2,395 5 Chestnut, R.W., Colon, S.M and Grey, H.M. (1982) J. Immunol. 129, 2382 6 Ishizaka, K., Okudaira, H. and King, T.P. (1975)J. ImmunoL 114, 110 7 Allen, P.M. and Unanue, E.R.(1984) J. Immunol. 132, 1077 $ Shimonkevitz, R., Colon, S., Kappler, J.W. etaL (1984)J. ImmunoL 133, 2067 9 Streicher, H.Z., Berkower, I.J., Busch, M. etal. (1984) Proc. Natl Acad. Sci. USA 81, 6831 10 EIIner,J.J., Lipsky, P.E.and Rosenthal,A.S. (1977) J. IrnmunoL 118, 2053 11 Lee, K.C., Wong, M and Spitzer, D. (1982) Transplantation 34, 150 12 Shimonkevitz, R., Kappler, J., Marrack, P. etaL (1983) J. Exp. Med. 158, 303 13 Castein, L.A., Lakey, E.K., Jelachich, M.L. etaL (1985) Proc. Natl Acad. Sci. USA 82, 5890 14 Chestnut, R.W., Colon, S.M. and Grey, H.M. (1982) J. ImmunoL 128, 1764 15 Kim, K.-H., Solvay, MJ. and Thomas, D.W. (1985) Cell. ImmunoL 96, 267 16 Kaye, P.M., Chain, B.M. and Feldman, M. (1985) J. ImmunoL 134, 1930 17 Kapsenberg, M.L., Teunissen,M.B.M., Stiekema, F.E.M. et al. (1986) Eur. J. ImmunoL 16, 345 Conclusions 18 Reske-Kunz, A.B., Reske, K. and Rude, E. (1986)J. Immunol. The nature and extent of the events which occur 136, 2033 between binding of an antigen to an APC surface 19 Walden, P., Nagy, Z.A. and Klein, J. (1986)J. Mol. Cell. membrane and recognition of an antigen determinant by ImmunoL 2, 191 class II-restricted T cells are clearly diverse and cannot be 20 Gotdberg, A.L. and St. John, A.C. (1976)Annu. Rev. Biochem. 45, 747 simply defined. The processing of a particular antigenic 21 Tietze, C., Schlesinger, P. and Stah[, P. (1980)Biochem. determinant appears to be affected by: (1) the conBiophys. Res. Commun. 93, 1 formation and size of the antigen; (2) the specificity of 22 Nowell, J. and Quaranta, V. (1985)J. Exp. Med. 162, 1371 the T cell, which determines the structure and location of 23 Malek, T.R., Clement, L.T. and Shevach, E.M (1983) Eur. J. the antigenic determinant; and/or (3) the properties of ImmunoL 13,810 the APC employed, their expression of specific surface 24 Mills, K.H.G., Skehel,J.J. and Thomas, D.B. (1986) Eur. J. receptors or class II molecules or their capacity for ImmunoL 16, 276 endocytosis and lysosomal proteolysis. In order to explain 25 Rott, R., Orlich, M, Klenk, H.-D. etal. (1984)EMBOJ. 3, the result of experiments in vitro alternative processing 2329 routes may be envisaged. One mechanism involves 26 Daniels, R.S., Douglas, A.R., Skehe[, J.J. etaL (1983)J. Gen. ViroL 64, 1657 endocytosis and proteolytic digestion while another 27 Scala, G. and Oppenheim, J.J. (1983)J. Immunol. 131, 1160 appears to bypass these events and allows recognition of 28 Buss,S. and Werdelin, O. (1986)J. ImmunoL 136, 459 tertiary structures. Degradative processing of globular 29 Mills, K.H.G., Skehel, J.J. and Thomas, D.B. (1986)J. Exp. antigen may be necessary to expose internal determiMed. 163, 1477 nants normally inaccessible to T cells, but may not be 30 Rock, K.L., Benacerraf, B. and Abbas, A.K. (1984)J. Exp. required for recognition of exposed conformational deMed. 160, 1102 terminants on the surface of the native antigen. 31 Morein, B., Barz, D., Koszinowski, U. etaL (1979) J. Exp. Med. 150, 1383 32 Parham, P. (1984) ImmunoL Today 5, 89-92 I wish to thank Brian Thomas, John Skehel and Ita Askonas 33 Grey, H.M., Colon, S.M. and Chestnut, R.W. (1982) J. ImmunoL 129, 2389 for helpful discussions.
lysosomal system is not essential for presentation of all globular antigens s'~l'ls'17'18, i.e. that non-lysosomal antigen processing can occur. Internalization of antigen may not be a prerequisite for antigen presentation and T cells may bind to antigen after a membrane-associated processing event 16'17'23. In addition, studies on class I-restricted T-cell recognition of virus proteins 3~'32 or class II-restricted T-cell recognition of alloantigen 27'32 have shown that T cells can see native antigen either unprocessed or re-expressed on the cell surface following virus infection of the APC. Further convincing evidence for alternative routes of antigen processing comes from recent reports suggesting that different APC use different pathways to process the same antigenic determinant. The results of one study ~s suggested that a macrophage cell line processed PPD (purified protein derivative) by an internal lysosomal route, whereas a B-cell line employed a membranebound mechanism for the same antigen. Similarly, Grey and coworkers 33 showed major differences between macrophages and B-cell processing of keyhole limpet hemocyanin (KLH); they found no evidence of lysosomal degradation or compartmentalization by B cells. Another report 17 suggested that, in contrast to macrophages dendritic cells do not internally degrade KLH or ovalbumin prior to presentation.
263
-rostrum
Processing of viral antigens and presentation to class II-restricted T cells Some antigens require intracellular processing by antigen presenting cells before being presented to T cells in conjunction with surface major histocompatibility complex antigens. The whole mechanism of these processing events is not known and in this article, Kingston Mills puts forward arguments for alternative routes of antigen processing, with particular reference to recognition of viral proteins by class II-restricted T-cell clones.
Activation of antigen-specific helper T (Th) cells requires recognition of the foreign antigen in association with class II major histocompatibility complex (MHC) gene products on the surface of an antigen presenting cell (APC) 1'2. Certain foreign macromolecules may require internal handling or processing by the APC before presentation of the antigenic determinant to the T cell 3-5. If we define antigen processing as the cellular and molecular events which occur between antigen binding to APC and subsequent recognition by the T cell, then it is evident from the literature that the nature of these events is uncertain, in particular the extent of biochemical change in the antigen.
Lysosomal antigen processing Unlike B cells which recognize determinants on the tertiary structure of a protein, class II-restricted T cells are thought to see mainly linear determinants on the primary structure of peptides, protein fragments or denatured antigen 6~1°. There is indirect evidence that large globular antigens require processing steps which involve denaturation and proteolysis in the lysosomal compartment of the APC: (1) a lag period, where metabolic events occur, exists between antigen binding to the APC and recognition by the T cell3,S; (2) antigens are internalized and transported to compartments within the APC which are resistant to protease strippingS'l°; (3) agents which inhibit lysosomal function by raising the pH (chloroquine and ammonia) or inhibit proteolytic enzymes (leupeptin) prevent presentation of globular antigens but not fragments or peptides7'9,11; and (4) aldehyde fixation of APC inhibits presentation of globular antigen but not peptides 7'9'12'13. There is increasing experimental evidence which challenges this original concept of antigen processing. For example, it is now clear that a wide variety of cells other than phagocytic macrophages can present antigen: resting and immune B cells 1 3,1 4 , B-cell lines or tumours 5,1 3 - 15, dendritic cells 16,17 and la + T cells 19 have all been shown to function as APC. Many of these cells either are not phagocytic, have poor endocytotic abilities, or have few lysosomes. Membrane preparations or liposomes containing the appropriate MHC protein and antigen are also capable of activating class II-restricted T cells ls'19 Experiments with the lysosomotropic agents have shown that these drugs do not always inhibit presentation of 260
NationalInstitutefor MedicalResearch,LondonNW7 1AA, UK
KingstonH.G. Mills globular antigens 5,11'~5'17. Furthermore, chloroquine or ammonia can inhibit other cell functions such as membrane receptor recycling or biosynthesis of la molecules 2c~22. Studies with radiolabelled antigens 23 have also suggested that a proportion of native antigen is not internalized but remains on the surface of the APC, where it could then be recognized by T cells after association with la.
Recognition of native structure by influenza virus-specific T-cell clones We have recently provided evidence that certain class II-restricted T cells can recognize conformational features of a native antigen 24. T-cell recognition was studied using influenza virus-specific Th-cell clones established from individual mice primed 3-6 months earlier by infection with X31 influenza virus. From a panel of clones specific for either the haemagglutinin (HA) surface glycoprotein or the internal matrix protein (MP), seven distinct HA-specific clones failed to respond to an X31 mutant virus (R19) with a single amino acid substitution (His to Arg) at position 17 of HA1 (Fig. 1)2s. Each of the seven clones also failed to recognize pH5-treated virus, a synthetic peptide corresponding to residues 1-27 of HA1 or tryptic fragments consisting of HA1 residues 28-328 (tops) and the remainder of the virion including residues 1-27 of HA1 (aggregate). A further twenty HA-specific clones responded to R19 and to tops and showed a positive or slightly reduced response to pH5-treated virus, whereas two MP-specific clones recognized R19, aggregate and pH5-treated virus. It is unlikely that the R19 negative clones are specific for an epitope in the region of position 17 since aggregate and peptide 1-27 were not recognized. It could be argued that the His to Arg substitution in R19 interferes with the ability of the H-2 k APC to process the antigen, or of the class II molecules to associate with an altered la interaction site. However, since H-2 k APC could present R19 to other influenza virus-specific H-2 k class IIrestricted clones, this explanation seems unlikely. A study of the structure of the R19 mutant virus suggested that the single amino acid substitution at position 17 affects the conformational stability of the HA molecule 25. By testing the specificity of the clones for natural variant viruses of known amino acid sequence, the recognition site of five of the seven clones was tentatively mapped to a determinant in the interface of the HA trimer in the region of residues 213 and/or 208. The interface antibody combining sites are antigenically and structurally altered in tops and in pH5-treated virus as a result of an irreversible conformational change which affects the integrity of the globular head region of the HA trimer 26. Therefore it was concluded that at least a proportion of class II-restricted T cells recognize conformational determinants on the three dimensional structure of the native HA molecule. This suggests that an alternative to the lysosomal degradative route of the ~) 1986,ElsevierSciencePublishe~B.V.,Arnsterdam 0167- 4919/86/$02.00
Immunology Today, voL 7, No. 9, 1986
ros/rum
Interface Antibody Combining Site H A Trimer
HA1
R19 (X31 M u t a n t Virus)
-- HA2
N a t i v e Virus + Chlor o q u i n e / A m m o n i a
/Leupeptin
~Viral N e t i v e X31 Virus
p H 5 Virus
Aggregate
H A - 1 Tops
Conformational-dependent H A - s p e c i f i c Clones
"1"
Other
-~,
,I- / - -
--
~)-
"~"
"0"
4"
~1"
--
"1"
HA Specific Clones
Internal P r o t e i n specific Clones
.
.
.
.
"~"
4"
Fig.1. Recognitionof viralproteinsby influenzavirus-specificclassII-restrictedT-cellclones. Theproliferativeresponseof a panelof T-cellcloneswas testedwith an optimumconcentrationof nativevirus,pHS-treatedvirus, viraltryptic fragments(HAI tops and aggregate)and an X31 mutant virus (RIg). Cloneswere divided into three groups. internal MP-specific; conformationaldependantHA-specific; and other HA-specificcloneswith a rangeof spedfidtiesinduding thosewhich recognizesyntheticpeptidesof HA29.Alsoshownare the responsesto the nativevirususingAPC pre-incubatedfor 1 h with chloroquine(0. I mM), or pre-incubatedand pulsedwith antigenfor 2 h in the presenceof ammoniumchloride(1.0 mM) or leupeptin(1.0 mM).
antigen processing does exist for certain HA-specific T cells. The low pH of the endosome or lysosome alone would alter the antigenic determinants recognized by these T cells. Therefore a route which can circumvent irreversible pH5-induced changes in structure and lysosomal proteolysis must be envisaged.
Effect of APC fixation on antigen presentation In an attempt to define possible mechanisms of antigen processing for HA- or MP-specific Th-cell clones, we tested the effect of cell fixation on antigen presentation. We failed to demonstrate presentation of native virus or BHA to any of the T-cell clones tested using glutaraldehyde or paraformaldehyde fixed splenic APC (unpublished). Proliferation was induced in a T-cell line, but only when the APC were pulsed with antigen before fixing. Furthermore, T-cell clones specific for a peptide of HA1 (48~8) failed to respond to the peptide with fixed APC. It is unlikely that a peptide as short as 20 amino acids would require further degradation for recognition. The fixed cells were not toxic since they did not inhibit T-cell proliferation in the presence of non-fixed APC. It is possible that lack of presentation by fixed cells may reflect a requirement for a feeder effect or IL-1 release by metabolically active APC27 or it may be due to distortion of the T-cell receptor or antigen interaction site on the la molecule. The majority of previous studies that demonstrated recognition of peptide antigens by fixed APC have involved measurement of IL-2 release by T-cell hybridomas as an index of T-cell activation. The requirements for induction of proliferation by T-cell clones may be more stringent.
Effect of lysosomotropic agents Experiments with lysosomotropic agents did provide evidence for alternative routes of antigen processing (unpublished). Chloroquine, ammonium chloride and leupeptin all inhibited presentation of whole virus or of purified MP to MP-specific clones. However, the effects on HA-specific clones was variable. Recognition of peptide 48-68, purified HA or native virus by peptidespecific clones was not inhibited to any great extent and in some cases (responses to virus or HA in presence of
leupeptin) was augmented. Residues 48-68 are on an exposed antibody combining region of the HA molecule, therefore it is possible that the T-cell clones may see a determinant on the unprocessed native virus, the structure of which can be adopted by the free synthetic peptide. Augmentation of proliferation in the presence of proteolytic inhibitors has previously been suggested to result from preservation of antigenic determinants in the native structure which might be lost during proteolysis 28. In the case of the conformational-dependent clones there was slight (10-40%) inhibition of proliferation to whole virus in the presence of chloroquine and leupeptin. This may argue for a lysosomal route of processing; however chloroquine may also affect the alternative routes of antigen processing described below.
Possible mechanisms for processing of influenza virus antigens It is evident from the findings with influenza virusspecific Th clones that more than one route exists for processing of viral antigens. Four possible alternatives are schematically described in Fig. 2. The first step in each route involves binding of the viral HA to a receptor on the APC membrane. This receptor may be a sialic acid containing membrane receptor or a specific anti-HA antibody on the surface of a B cell. In the case of exogenous peptide antigen, binding may simply involve non-specific sticking to the APC membrane. Route A is the classical lysosomal route and involves endocytosis, fusion of the endosome with a lysosome, denaturation and proteolytic digestion in the acidic environment of the lysosome and excretion of fragments to the cell surface. This route would be blocked by lysosomotropic agents and may be the one used for processing of internal viral proteins. However, it is just as likely that an alternative internal pathway is employed (route B). In this case the contents of the endosome are dissociated by a mechanism which does not require entry into a lysosome. This mechanism may involve release of the viral contents into the cytoplasm. Non-lysosomal pathways of proteolysis, such as the ubiquitin-associated system, have been described for several proteins and may also be blocked by chloroquine or ammonium chloride through their effect on endocytosis2°
261
Immunology Today, vol. 7, No. 9, 1986
rostru I Antigen Binding to Specific Receptor
lI
111
Endocytosis
I~
Denaturation and Proteolysis
Exocytosis
~" Association with la
31I Ag-la-TcR Ternary Complex Formation
Exogenous Antigen Fragments Viral
T Cell
Lysosomal Degradation Endosome
T Dell
Lysosom~ NP
Native Virion
# Cytoplasmic
Fusion of Viral and
Native HA Molecule
'Membrane Processing'
(~
Exogenous BHA
Fig.2. Schematicrepresentationof alternativeroutesfor processingof influenzavirusantigen. Fourpossibleroutes(A-D)areshownwithsixindividualsteps(I-VI)for routeA, manyof whicharebypassedin the otherroutes.A: internallysosomalprocessing.B: non-lysosomalinternaldegradation.C: infectionof the APCby the virusfollowedbyexpressionof nativeglycoproteinson the cellsurfacemembrane.D: presentation of nativeantigenwithoutinternalization.
Route C requires infection of the APC by the virus. During infection fusion of viral and endosomal membranes occurs permitting release of the viral RNA for transcription and translation. Newly synthesized viral glycoproteins are excreted via the Golgi apparatus and expressed on the external surface of the cell membrane, where they could then be recognized by the T cell in their native conformation. The internal pathways are bypassed in route D. Here native virus or HA is presented to the T-cell without internal processing. Certain antigens may undergo proteolysis at the cell surface membrane 28 ; however this is unlikely to occur with the relatively proteolytic-resistant native influenza virion. Therefore the only step between virus binding to the cell surface receptor and T-cell recognition may involve association of the antigen with class II molecules. This association may be a loose interaction which is stabilized after binding of the T-cell receptor or it may only occur in the presence of the T-cell receptor.
262
Evidence for alternative routes of antigen processing We may conclude from the studies with influenza virus-specific Th-cell clones that internalization and lysosomal degradation of globular antigens are not essential features of antigen processing. Alternative mechanisms which permit recognition of native structures by T cells clearly exist. Interestingly, recognition of the conserved MP, which is enclosed within the viral membrane,
appears to require an intracellular proteolytic processing step, whereas the exposed HA molecule can be recognized in its native shape. Therefore the specificity of the T cell may determine the requirement for antigen processing. The previous assumption that class II-restricted T cells recognize linear determinants on processed antigen was based on studies with conserved proteins and with T cells derived from mice immunized with antigen in adjuvant. We have used T-cell clones established from mice primed by infection to demonstrate a previously unrecognized diversity in class II-restricted T-cell recognition of antigen, including specificity for determinants in exposed variable antibody combining regions of the HA molecule 29. Aggregation of an antigen with adjuvant, which is required for effective priming with soluble globular antigens, renders it less susceptible to recognition in its native conformation. Immunization by infection overcomes the requirement for priming in adjuvant and may explain previous failures to detect Th-cell recognition of tertiary structures. Furthermore memory T cells may employ a different mechanism for antigen processing than that required for a primary response. Antigenspecific B cells may focus antigen in vivo, via their immunoglobulin receptors, for recognition by antigenspecific T cells. Immune B cells have been shown to function as effective APC in vitro at lower antigen concentration than that required for macrophages 14'3°. A number of studies have shown that a functional
ImmunologyToday,vol. 7, No. 9, 1986
ros/rum
References 1 Kappler, J.W. and Marrack, P.C.(1976) Nature (London) 262, 797 2 Schwartz, R.H. (1985)Annu. Rev. Immunol. 3, 237 3 Ziegler, K. and Unanue, E.R.(1982) J. ImmunoL 127, 1869 4 Unanue, E.R.(1982)Annu. Rev. ImmunoL 2,395 5 Chestnut, R.W., Colon, S.M and Grey, H.M. (1982) J. Immunol. 129, 2382 6 Ishizaka, K., Okudaira, H. and King, T.P. (1975)J. ImmunoL 114, 110 7 Allen, P.M. and Unanue, E.R.(1984) J. Immunol. 132, 1077 $ Shimonkevitz, R., Colon, S., Kappler, J.W. etaL (1984)J. ImmunoL 133, 2067 9 Streicher, H.Z., Berkower, I.J., Busch, M. etal. (1984) Proc. Natl Acad. Sci. USA 81, 6831 10 EIIner,J.J., Lipsky, P.E.and Rosenthal,A.S. (1977) J. IrnmunoL 118, 2053 11 Lee, K.C., Wong, M and Spitzer, D. (1982) Transplantation 34, 150 12 Shimonkevitz, R., Kappler, J., Marrack, P. etaL (1983) J. Exp. Med. 158, 303 13 Castein, L.A., Lakey, E.K., Jelachich, M.L. etaL (1985) Proc. Natl Acad. Sci. USA 82, 5890 14 Chestnut, R.W., Colon, S.M. and Grey, H.M. (1982) J. ImmunoL 128, 1764 15 Kim, K.-H., Solvay, MJ. and Thomas, D.W. (1985) Cell. ImmunoL 96, 267 16 Kaye, P.M., Chain, B.M. and Feldman, M. (1985) J. ImmunoL 134, 1930 17 Kapsenberg, M.L., Teunissen,M.B.M., Stiekema, F.E.M. et al. (1986) Eur. J. ImmunoL 16, 345 Conclusions 18 Reske-Kunz, A.B., Reske, K. and Rude, E. (1986)J. Immunol. The nature and extent of the events which occur 136, 2033 between binding of an antigen to an APC surface 19 Walden, P., Nagy, Z.A. and Klein, J. (1986)J. Mol. Cell. membrane and recognition of an antigen determinant by ImmunoL 2, 191 class II-restricted T cells are clearly diverse and cannot be 20 Gotdberg, A.L. and St. John, A.C. (1976)Annu. Rev. Biochem. 45, 747 simply defined. The processing of a particular antigenic 21 Tietze, C., Schlesinger, P. and Stah[, P. (1980)Biochem. determinant appears to be affected by: (1) the conBiophys. Res. Commun. 93, 1 formation and size of the antigen; (2) the specificity of 22 Nowell, J. and Quaranta, V. (1985)J. Exp. Med. 162, 1371 the T cell, which determines the structure and location of 23 Malek, T.R., Clement, L.T. and Shevach, E.M (1983) Eur. J. the antigenic determinant; and/or (3) the properties of ImmunoL 13,810 the APC employed, their expression of specific surface 24 Mills, K.H.G., Skehel,J.J. and Thomas, D.B. (1986) Eur. J. receptors or class II molecules or their capacity for ImmunoL 16, 276 endocytosis and lysosomal proteolysis. In order to explain 25 Rott, R., Orlich, M, Klenk, H.-D. etal. (1984)EMBOJ. 3, the result of experiments in vitro alternative processing 2329 routes may be envisaged. One mechanism involves 26 Daniels, R.S., Douglas, A.R., Skehe[, J.J. etaL (1983)J. Gen. ViroL 64, 1657 endocytosis and proteolytic digestion while another 27 Scala, G. and Oppenheim, J.J. (1983)J. Immunol. 131, 1160 appears to bypass these events and allows recognition of 28 Buss,S. and Werdelin, O. (1986)J. ImmunoL 136, 459 tertiary structures. Degradative processing of globular 29 Mills, K.H.G., Skehel, J.J. and Thomas, D.B. (1986)J. Exp. antigen may be necessary to expose internal determiMed. 163, 1477 nants normally inaccessible to T cells, but may not be 30 Rock, K.L., Benacerraf, B. and Abbas, A.K. (1984)J. Exp. required for recognition of exposed conformational deMed. 160, 1102 terminants on the surface of the native antigen. 31 Morein, B., Barz, D., Koszinowski, U. etaL (1979) J. Exp. Med. 150, 1383 32 Parham, P. (1984) ImmunoL Today 5, 89-92 I wish to thank Brian Thomas, John Skehel and Ita Askonas 33 Grey, H.M., Colon, S.M. and Chestnut, R.W. (1982) J. ImmunoL 129, 2389 for helpful discussions.
lysosomal system is not essential for presentation of all globular antigens s'~l'ls'17'18, i.e. that non-lysosomal antigen processing can occur. Internalization of antigen may not be a prerequisite for antigen presentation and T cells may bind to antigen after a membrane-associated processing event 16'17'23. In addition, studies on class I-restricted T-cell recognition of virus proteins 3~'32 or class II-restricted T-cell recognition of alloantigen 27'32 have shown that T cells can see native antigen either unprocessed or re-expressed on the cell surface following virus infection of the APC. Further convincing evidence for alternative routes of antigen processing comes from recent reports suggesting that different APC use different pathways to process the same antigenic determinant. The results of one study ~s suggested that a macrophage cell line processed PPD (purified protein derivative) by an internal lysosomal route, whereas a B-cell line employed a membranebound mechanism for the same antigen. Similarly, Grey and coworkers 33 showed major differences between macrophages and B-cell processing of keyhole limpet hemocyanin (KLH); they found no evidence of lysosomal degradation or compartmentalization by B cells. Another report 17 suggested that, in contrast to macrophages dendritic cells do not internally degrade KLH or ovalbumin prior to presentation.
263
