Proc. Natl. Acad. Sci. USA Vol. 85, pp. 1209-1212, February 1988 Immunology Antigen presentation to HLA class II-restricted measles virus-specific T-cell clones can occur in the absence of the invariant chain (antigen processing/chloroquine/DNA-mediated gene transfer/HLA-DR restriction/major histocompatibility complex) RAFICK P. SEKALY*t, STEVEN JACOBSONt, JOHN R. RICHERT§, CECILE TONNELLE*¶, HENRY F. MCFARLANDt, AND ERIC 0. LONG* *Laboratory of Immunogenetics, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892; tNeuroimmunology Branch, National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892; and §Department of Neurology, Georgetown University Medical Center, Washington, DC 20007 Communicated by William E. Paul, November 9, 1987

ABSTRACT A human fibroblast line expressing HLA- The requirements for cell-surface expression of class II a/3 DR1 antigen on its surface was generated by transfection with heterodimers have also been analyzed by DNA-mediated DRa and DRP cDNAs. Expression of the invariant chain gene gene transfer. It was recently shown that the invariant chain was not detectable in the transfected fibroblasts and was not that is transiently associated with a,8 heterodimers intracel- induced by infection with measles virus. Lysis of measles lularly is not required for expression of human MHC class II virus-infected cells occurred with DRM- but not with DR4- antigens at the cell surface (8, 9). These results suggested restricted measles virus-specific cytotoxic T-lymphocyte (CTL) that the invariant chain might be involved in the function of clones and was inhibited by a monoclonal antibody specific for MHC class II antigens, such as binding of processed antigen DR antigen. Therefore, the invariant chain is not required for fragments, rather than in their biosynthesis. In this report, DR-restricted presentation of measles virus antigens by this we have used clones of a transfected human fibroblast line fibroblast line. Transfected fibroblasts were lysed as efficiently that expressed DR antigens in the absence of the invariant as an autologous B-cell line even though they expressed much chain to test for their ability to present antigens. Presenta- less surface DR antigen. Lysis of both the transfected fibro- tion of measles virus antigens to specific CTLs occurred blasts and the B-cell line was insensitive to treatment with efficiently in these cells. chloroquine. These results demonstrate that expression of a DR a'p heterodimer at the surface of this fibroblast line is MATERIALS AND METHODS necessary and sufficient for presentation of measles virus antigens to specific CTL clones. Transfected Human Fibroblast Lines. The human simian virus 40-transformed fibroblast line 637B and two clones The interaction of a helper T cell with an antigen-presenting obtained after transfection with the DR1 a- and 8-chain cell that expresses surface class II antigens and a specific cDNAs have been described (9). Briefly, 637B cells were antigen results in lymphokine production by the T cell. transfected by the calcium phosphate-precipitation tech- Lymphokines in turn trigger effector cells of the immune nique with plasmids containing full-length DRa and DRP system (1). Human antigen-presenting cells (B cells, macro- cDNA clones together with the pSV2-neo plasmid for selec- phages) express class II antigens of the major histocompat- tion. Expression of cDNA inserts was controlled by the ibility complex (MHC)-namely, HLA-DR, -DQ, and -DP simian virus 40 early promoter. DR-positive 637B cells were (2). Recent advances in cloning the genes for the a and /3 isolated after selection in G418 by several cycles of cell subunits of class II antigens and their expression in cells that sorting and then were cloned (9). are normally class II negative have made it possible to study Viral Infections. Target cells were infected with the Ed- the structure-function relationships in class II antigens. monston strain of measles virus (6, 7) at a multiplicity of Murine fibroblasts transfected with the genes for murine or infection of 1 [for the B lymphoblastoid cell line (B-LCL)] or human MHC class II antigens can express the molecules on 3 (for 637B) for 90 min and incubated for 3 days in Eagle's the cell surface (3-5). Moreover, these class II-expressing minimal essential medium supplemented with 15% heat- transfectants are capable of stimulating several antigen- inactivated fetal calf serum and 2 mM glutamine. Target cells specific and allospecific T cells in a MHC-restricted manner were infected with influenza virus as described (7). (4, 5). RNA Blot Analysis. Poly(A) + RNA from the untransfected MHC class II antigens can also be involved in antigen 637B cell line, transfected cell lines, and control B-cell line presentation to cytotoxic T lymphocytes (CTLs). The CTL were denatured with glyoxal, size-fractionated in 1.5% aga- response to measles virus is predominantly MHC class II rose gels, and transferred to nylon membranes as described restricted and the majority of measles virus-specific CTL (9). Prehybridizations and hybridizations with DNA frag- cones are CD4+ (6, 7). To analyze the requirements for the ments from DRa, DR/3, or invariant chain cDNAs were recognition of measles virus-infected cells by CTLs, DNA- carried out as described (9). mediated gene transfer of human MHC class II genes was Effector Cells. The measles virus-specific CTL clones used to reconstitute a suitable target cell. In this report, we were generated and maintained in culture as described (6, 7). show that expression of HLA-DR at the surface of a human Clones TB30, TB34, TB38, and TB60 were shown to be fibroblast line infected with measles virus is sufficient for recognition and lysis by specific CTLs. Abbreviations: B-LCL, B lymphoblastoid cell line; CTL, cytotoxic T lymphocyte; MHC, major histocompatibility complex. tPresent address: Clinical Research Institute of Montreal, Montreal, The publication costs of this article were defrayed in part by page charge Quebec H2W1R7, Canada. payment. This article must therefore be hereby marked "advertisement" Present address: Centre d'Immunologie de Marseille-Luminy, in accordance with 18 U.S.C. §1734 solely to indicate this fact. 13288 Marseille, France.

Downloaded by guest on September 27, 2021 1209 1210 Immunology: Sekaly et al. Proc. Natl. Acad. Sci. USA 85 (1988) DR1-restricted by their ability to lyse a panel of DR1- transfected 637B lines was analyzed. The RNA transcripts positive B-LCL infected with measles virus and their inabil- encoded by the cDNA expression vector are =300 bases ity to lyse infected DR1-negative B-LCL. The DR4- longer than the natural transcripts because they contain restricted measles virus-specific CTLs were generated by additional vector sequences (9, 14). With this size difference, primary in vitro sensitization from a normal donor as de- it was possible to determine that measles virus infection did scribed (7). These CTLs cultured in bulk were shown to be not induce expression of the endogenous DRa and DRf3 predominantly CD4+ and HLA class II restricted (7, 10, 11). genes (Fig. 1). Longer transcripts from the transfected genes The DR4-restricted influenza virus-specific CTL line was were also detected; they most likely result from transcription generated by repeated stimulations with irradiated (6000 rad; through the simian virus 40 late gene transcription termina- 1 rad = 0.01 Gy) autologous influenza virus-infected periph- tion signal present on the vector and termination at down- eral blood lymphocytes as feeders and cultured in the stream flanking sequences. presence of 10% interleukin 2 and 5% human AB serum. The invariant chain gene is not expressed in 637B cells but After seven passages, this line expressed the CD4' pheno- it can be induced, together with other class II genes, by type and lysed specifically DR4-positive B-LCL targets treatment with y-interferon (9). It is also known that viral infected with influenza virus. particles can induce class II antigen expression on certain CTL Assays. An indium oxine ("1'In) release assay (12) cells (15). It was therefore important to test whether the was used with 637B cells because it offered a substantially 3-day infection with measles virus had induced expression of improved efficiency of labeling over 51Cr. Microtiter wells the invariant chain gene in 637B cells. No transcript from the containing confluent monolayers of transfected or untrans- invariant chain gene was detected in fected cells were washed twice with complete Hanks' bal- infected 637B-C9 cells anced salt solution and incubated with 200 jul of '11In at a (Fig. 1). Note that 5 gg of 637B-C9 poly(A)+ RNA was run final concentration of 2 gCi per well (1 Ci = 37 GBq), for 15 in parallel with only 0.5 ,ug of poly(A)+ RNA from a B-LCL. min at room temperature and then washed twice with Hanks' Antigen Presentation to Measles Virus-Speciflc DR1- balanced salt solution supplemented with bovine serum Restricted CTL Clones Occurs in the Absence of the Invariant albumin (40 mg/ml) fraction V. After several washes, the Chain. The ability of transfectants that lack invariant chain microwells were incubated in Hanks' balanced salt solu- expression to present antigens to measles virus-specific tion/bovine serum albumin for 90 min at 370C. Effector cells DR1-restricted CTL clones was tested. Results (Table 1) were added and incubated for 6 hr with the target cells. indicated that no specific lysis occurred when untransfected Supernatants were harvested and released 1"'in was mea- 637B cells were used as targets even when infected with sured by using a Micromedicine Plus model y-counter measles virus. Similar results were obtained with a G418- (Micromedic Systems, Horsham, PA). Percentage specific resistant DR-negative clone obtained from the same trans- release was calculated as follows: (cpm of experimental fection (data not shown). Furthermore, the uninfected DR1- release - cpm of spontaneous release)/(cpm of total release positive C9 transfectant was not recognized by the CTL - cpm of spontaneous release). In most experiments, spon- clones. In contrast, C9 cells infected with measles virus were taneous release varied between 15% and 20%. Total release lysed by several DR1-restricted measles virus-specific CTL was obtained by treating the "1In-labeled cells with 1% clones. Background levels of lysis were observed with a Triton X-100. CTL assays with B-LCL were performed with DR4-restricted CTL clone, confirming that killing of the 51Cr as described (6, 7). Chloroquine inhibition experiments target cells was DR1 restricted. involving influenza virus were performed according to Mor- Further evidence that measles virus antigens were recog- rison et al. (13). Chloroquine inhibition experiments involv- nized by the CTL clones in the context of DR was obtained ing measles virus were modified by washing 637B fibroblasts by blocking experiments with monoclonal antibodies to class or B-cell targets with 50 AM chloroquine prior to infection I and class II antigens. The measles virus-infected C9 with measles virus. During viral infection, cells were cul- transfectant was incubated with monoclonal antibodies for tured for 3 days in the presence of 50,uM chloroquine. Viral 637B 637B 637B antigen expression was identical in cells treated with or I without chloroquine. Cells were harvested and "1'In-labeled > > + LUi + LU V + l.LV a, 0% 0% Ln 0% 0%l 0% Ln 0% o% c7% Lm 637B cells or 51Cr-labeled B-LCL targets were diluted to 10 Li Li Lij -~t Lij Li Li -* kLi ti Li -* ,uM chloroquine and a 1:1 dilution of target cells was made with effectors yielding a final chloroquine concentration of 5 ,M in the 6-hr CTL assay. 18S- 3 RESULTS id w Measles Virus Infection of a Human Fibroblast Line That e* Expresses Transfected DRa and DRP cDNAs. The human fibroblast line 637B was infected in vitro with measles virus. After infection, 637B cells were still negative for expression DRor DR 13 In of MHC class II antigens (data not shown). The 637B line FIG. 1. RNA blot analysis of transfected and untransfected has been previously transfected with plasmids carrying cells. Poly(A)+ RNA (5 jug) from the 637B cells and 0.5 ,ug of expressible cDNA inserts for the DRa and DR.8 chains (9). poly(A)+ RNA from the B-cell line 45.1 mixed with 4.5 j&g of total The 637B cells expressing cell-surface DR antigen were HeLa cell RNA as carrier (9) were size-fractionated by gel electro- isolated by cell sorting on a flow cytometer and then were phoresis and transferred to a nylon membrane. Lanes: C9, a clone of cloned. The level of surface DR antigen on clones C9 and 637B cells transfected with HLA-DR genes; C9 + MV, C9 cells C9E8 was about 1/20th to 1/10th the level present on a infected with measles virus; C9E8, a subclone of C9 that expresses B-LCL (9). DR expression in transfected 637B cells was higher levels of DRi antigen; 45.1, the HLA-hemizygous B-cell line unaffected by measles virus infection (data not shown). from which the cDNA clones were derived (14). The filters were Antibodies for the other MHC class II antigens DQ hybridized with labeled DNA fragments from a DRa-chain cDNA specific (DRa), a DR,8-chain cDNA (DR,8), or an invariant chain cDNA(In). and DP did not bind to DR-positive 637B transfectants, even The additional band for DRa in C9 + MV cells is visible in all after measles virus infection. transfected 637B cells at longer exposures. The 637B cells were still To confirm that only the transfected class II genes were negative for In at longer exposures. The positions of 18S and 28S transcribed, poly(A)+ RNA from the transfected and un- RNA are indicated. Downloaded by guest on September 27, 2021 Immunology: Sekaly et al. Proc. Natl. Acad. Sci. USA 85 (1988) 1211 Table 1. Lysis of target cells by measles virus-specific CTL clones Target cell CTL clone B-1JCL 637B 637B-C9 (DR restriction) E/T - MV + MV - MV + MV - MV + MV TB-30 20:1 3.7 25.7 4.9 9.5 -0.5 32.8 (DR1) 7:1 4.1 16.3 -1.3 -3.1 -0.5 35.4 TB-34 8:1 2.8 33 NT NT -0.5 46.6 (DR1) 2:1 -3.0 22 NT NT -0.4 9.3 TB-38 13:1 3.0 29.1 3.4 6.4 -4.6 26.2 (DR1) 4:1 1.9 15.7 5.9 6.3 1.6 27.9 E4-1 20:1 1.9 24.5 14.8 8.5 11.4 15.4 (DR4) 7:1 -3.6 12.0 5.6 5.7 0.1 4.3 Each target cell was assayed before (-MV) or after (+MV) infection with measles virus. The B-LCL was typed as DR1, -4. NT, not tested. E/T, effector/target cell ratio. 30 min prior to the CTL assay. A DR-specific monoclonal to follow a pathway different from the endocytic pathway antibody inhibited lysis by two different CTL clones in a that is involved in processing of influenza virus antigens. dose-dependent manner (Fig. 2). In contrast, monoclonal antibodies to DP and to class I antigens had little effect. DISCUSSION Antigen Presentation to Measles Virus-Specific CTL Clones DNA-mediated gene transfer of human MHC class II genes Is Chloroquine Insensitive. Chloroquine is a lysosomotropic into a human class Il-negative fibroblast line was used to agent known to inhibit processing of soluble antigens or viral study the requirements for antigen recognition by measles antigens that are presented by MHC class II antigens (13, virus-specific CTL. Expression of DR antigen on the surface 16). A class II-restricted CD4' CTL line specific for influ- of the fibroblast line was both sufficient and necessary for enza virus was used as a control (Fig. 3A). As expected, lysis efficient recognition, after measles virus infection, by mea- of an influenza virus-infected B-LCL by this CTL line was sles virus-specific T-cell clones. It can also be concluded completely blocked by treatment of the B-LCL with chloro- that the invariant chain is not required in this process of quine prior to and during the infection. The chloroquine antigen presentation because expression of the invariant treatment blocked antigen processing but not infection be- chain gene was undetectable in the transfected fibroblasts cause high levels of influenza virus antigens were detected at even after infection with measles virus. Previous studies of the cell surface (data not shown). In contrast, the same antigen presentation by murine L cells transfected with class B-LCL infected with measles virus in the presence of II genes (4, 5) did not address the role of the invariant chain chloroquine was still lysed by measles virus-specific CTLs because L cells express this protein (17). (Fig. 3B). Two to 4 times higher effector/target cell ratios The transfected fibroblast line infected with measles virus were required to achieve the level of lysis seen with infected was lysed as efficiently as an infected B-cell line by measles cells not treated with chloroquine. The basis for this differ- virus-specific CTL clones even though it expressed 10 to 20 ence is either that the chloroquine-treated B-LCL is not times less DR antigen at the cell surface. This efficient lysed as easily as the untreated B-LCL, or that some effector recognition and lysis of a transfected fibroblast by human cells in the bulk measles virus-specific CTL population CTLs contrasts with several reports on transfected murine recognized antigens that are processed through a chloro- fibroblasts (18, 19). Even though transfected murine fibro- quine-sensitive pathway. The transfected fibroblasts in- blasts have been lysed by certain human CTLs (20), it is fected with measles virus in the presence of chloroquine likely that the use of a human target cell accounts for the were lysed by a DR1-restricted measles virus-specific CTL efficient lysis by human CTLs reported here. Antigen- clone as as cells independent conjugate formation between target and effec- efficiently not treated with chloroquine (Fig. tor cells may be essential for target cell recognition by CTLs 3C). Therefore, processing of measles virus antigens seems (21). It is relevant in this respect that the fibroblast line 637B expresses LFA-3, a surface molecule that contributes to A B conjugate formation by binding to the CD2 T-cell surface Q molecule (21, 22). 10 Several pathways of antigen presentation exist (23). Solu- ble extracellular proteins can be endocytosed by antigen- z 30 presenting cells and are presumably processed in endocytic 1- 40- vesicles. This exogenous pathway is typically linked to z antigen presentation to class II-restricted T cells, and it is Uu 50- blocked by chloroquine (16). The endogenous pathways of LU 60- antigen presentation have not yet been defined but are 70 involved in class I-restricted presentation of cytoplasmic and cell-surface viral antigens (24). This presentation of endog- 0 100 200 0 100 200 tg/ml enously synthesized viral antigens is chloroquine insensitive o o.1 0.2 0 0.1 0.2 DILUTION (13). Although no evidence exists yet for class II-restricted presentation of newly synthesized proteins, several findings FIG. 2. Antibody inhibition of lysis by measles virus-specific that the DRI-restricted CTL clones. Monoclonal antibodies specific for DR suggest processing of measles virus antigens in antigen (L243, purified antibody; *), DP antigen (B7.21, dilution of infected cells occurs via an endogenous pathway. First, as hybridoma supernatant; *), and class I antigens (W6.32, purified demonstrated in this study, the measles virus antigen pre- antibody; *) were incubated for 30 min with target cells prior to sentation is not blocked by treatment of target cells with incubation with clone TB60 (A) or clone TB38 (B). Targets were chloroquine. The lack of inhibition by chloroquine is not a measles virus-infected 637B-C9 cells. The effector/target cell ratio peculiar feature of the transfected cell expressing DR antigen was 10 and specific lysis in the absence of antibody was 40o. in the absence of the invariant chain, because it was also Downloaded by guest on September 27, 2021 1212 Immunology: Sekaly et al. Proc. Natl. Acad. Sci. USA 85 (1988)

Ak in endosomes, via the exogenous pathway. Newly synthe- Target B-LCL sized class II antigens, as a3-invariant chain complexes, do Influenza - + + intersect endocytic vesicles (26) and chloroquine interferes -- with the from af3 Chloroquine + release of the invariant chain the dimer (27). Class II-restricted CTL specific for influenza virus- 40 infected cells would be a good experimental system to test >a for the role of the invariant chain in the exogenous pathway 30. of antigen presentation. Class II-restricted CTL clones spe- cific for influenza virus have to be generated for that ._* 20- purpose, because bulk CTL lines give a high background killing with the 637B fibroblast line. ; 10 The results of this study demonstrate that measles virus- specific DR1-restricted CTLs require only the appropriate B class II antigen on the infected target cell for effective recognition. The use human cells Target B-LCL of for introduction of class II MHC genes has circumvented difficulties encountered in Measles other studies. Such a homologous system is also well suited to Chloroquine study the role of accessory molecules and antigen-processing - + +~~ ._ 40 pathways both at the effector and target cell level. - +~~~~~~~ , 1. Schwartz, R. H. (1985) Annu. Rev. Immunol 3, 237-261. 30- 2. Kaufman, J. F., Auffray, C., Korman, A. J., Schackelford, D. A. & Strominger, J. L. (1984) Cell 36, 1-13. u 20- X~~~~~ 3. Rabourdin-Combe, C. & Mach, B. (1983) Nature (London) 303, 670-674. H 10- 4. Malissen, B., Peele Price, M., Goverman, J., McMillan, M., /7 White, J. D., Kappler, J., Marrack, P., Pierres, A., Pierres, M. & Hood, L. 9. ./ (1984) Cell 36, 319-327. C 5. Norcross, M. A., Bentley, D. M., Margulies, D. H. & Ger- Target main, R. N. (1984) J. Exp. Med. 160, 1316-1337. 6. Richert, J. R., McFarland, H. F., McFarlin, D. E., Bellini, Measles W. J. & Lake, P. (1983) Proc. Natl. Acad. Sci. USA 80, Chloroquine 255-258. 40- 7. Jacobson, S., Richert, J. R., Biddison, W. E., Satinsky, A., Hartzman, R. J. & McFarland, H. F. (1984) J. Immunol. 133, 754-757. 30- 8. Miller, J. & Germain, R. N. (1986) J. Exp. Med. 164, 1478-1489. 20- 9. Sekaly, R. P., Tonnelle, C., Strubin, M., Mach, B. & Long, ak E. 0. (1986) J. Exp. Med. 164, 1490-1504. R 102 10. Jacobson, S., Fleralage, M. L. & McFarland, H. F. (1985) J. Exp. Med. 162, 839-850. 11. Jacobson, S., Rose, J. W., Fleralage, M. L., Mingioli, E. S., 2.5 10 40 2.510 40 2.5 10 40 McFarlin, D. E. & McFarland, H. F. (1987) in The Biology of Effector to target ratio Negative Strand Viruses, eds Kolakofsky, D. & Mahy, B. W. J. (Elsevier, Amsterdam), pp. 283-289. 12. Collins, T., Krensky, A. M., Gim- FIG. 3. Effect of chloroquine on the lysis of virus-infected cells Clayberger, C., Fiers, W., brone, M. S. J. & by virus-specific CTL. Target cells were either infected or A., Burakoff, Pober, J. S. (1984) J. Immu- not nol. 133, 1878-1884. infected with influenza virus or measles virus as indicated. Infected 13. Morrison, L. A., Lukacher, A. E., V. D. P. cells were either treated or not treated with chloroquine as indi- Braciale, L., Fan, & Braciale, T. J. (1986) J. Med. 903-921. cated. Uninfected cells treated with chloroquine gave background Exp. 163, 14. Tonnelle, C., DeMars, R. & Long, E. 0. (1985) EMBO J. 4, levels of lysis (data not shown). (A) DR4-restricted CD4' influenza 2839-2847. virus-specific CTLs were used as effectors. The B-LCL was HLA- matched to the only at DR. (B) 15. Massa, P. T., D6rries, R. & ter Meulen, V. (1986) Nature effectors DR4-restricted measles (London) 320, 543-546. virus-specific CTLs were used as effectors. Targets were an autol- 16. Unanue, E. R. (1984) Annu. Rev. Immunol. 2, 395-428. ogous B-LCL. (C) The DR1-restricted measles virus-specific CTL 17. Koch, N. & Harris, A. W. (1984) J. Immunol. 132, 12-15. clone TB38 was used as effector. Targets were 637B-C9 cells. 18. Bernabeu, C., Burakoff, S. J. & Terhorst, C. (1984) J. Immu- nol. 133, 3188-3194. observed with a B-LCL. we Second, have obtained lysis by 19. Barbosa, J. A., Mentzer, S. J., Minowada, G., Strominger, measles virus-specific CTLs of target cells transfected with J. L., Burakoff, S. J. & Biro, P. A. (1984) Proc. Natl. Acad. measles virus genes (unpublished data). Finally, unlike in- Sci. USA 81, 7549-7553. fluenza viruses, which fuse with membranes at the low pH 20. Cowan, E. P., Coligan, J. E. & Biddison, W. E. (1985) Proc. environment of endocytic vesicles, paramyxoviruses, such Natl. Acad. Sci. USA 82, 4490-4494. 21. Shaw, S., Luce, G. E., Quinones, R., R. as measles, fuse with the cell membrane at Gress, E., Springer, neutral pH (25). T. A. & Sanders, M. E. (1986) Nature (London) 323, 262-264. The primary source of measles antigen for processing and 22. Dustin, M. L., Sanders, M. E., Shaw, S. & Springer, T. A. presentation may thus be cytoplasmic proteins rather than (1987) J. Exp. Med. 165, 677-692. endocytosed material. 23. Germain, R. N. (1986) Nature (London) 322, 687-689. The lack of requirement for the invariant chain in antigen 24. Townsend, A. R. M. (1987) Immunol. Res. 6, 80-100. 25. Nagai, Y., Hamaguchi, M., Toyoda, T. & Yoshida, T. (1983) presentation by class II antigens may apply only to the Virology 130, 263-268. endogenous pathway of antigen processing. It is still possible 26. Cresswell, P. (1985) Proc. Natl. Acad. Sci. USA 82, that the invariant chain is required for proper antigen pre- 8188-8192. sentation by class II antigens when processing is taking place 27. Nowell, J. & Quarranta, V. (1985) J. Exp. Med. 162, 1371-1376. Downloaded by guest on September 27, 2021