Analysis of function of a human antigen-presenting cell by xenogeneic interaction with mouse T cells

Immunology Letters, 40 (1994) 73-77 Elsevier Science B.V. IMLET 02109

Analysis of function of a human antigen-presenting cell by xenogeneic interaction with mouse T cells Y a s u k o T s u n e t s u g u - Y o k o t a a'*, T o s h i a k i M i z u o c h i b, H i s a k o H a s h i m o t o a, A n d r e j S z i k a r a d k i e w i c z c, H i d e o Y a g i t a d, A k i h i k o Y a n o e a n d T o s h i t a d a T a k e m o r i a "Department of Immunology and bDepartrnent of Bacterial and Blood Products, National Institute of Health, Tokyo, Japan; Clnstitute of Microbiology and Infectious Diseases, Medical Academy, Poznan, Poland," aDepartment of lmmunology, Juntendo University School of Medicine, Tokyo, Japan; eDepartment of Experimental Animals, Nagasaki University School of Medicine, Nagasaki, Japan (Received 4 January 1994; accepted 8 February 1994)

1. Summary A human B cell line, ARH, was transfected with a murine major histocompatibility complex class II gene (I-Ak). One of the transfectants, ARH5.5, which strongly expresses I-A k molecules was found to be capable of presenting soluble antigens to I-A krestricted, antigen-specific murine helper T cell (Th) clones. When ARH5.5 was treated with either chloroquine or paraformaldehyde prior to the antigen pulse, it failed to present a protein antigen, ovalbumin, but retained the ability to present a peptide, indicating that the presentation was dependent on processing. The xenogeneic interaction of co-stimulatory molecules on the human antigen presenting cell (APC) and the murine Th cell was assessed by using antibodies against adhesion molecules. We found that the xenogeneic interaction of LFA-I/ICAM-1 acted as a strong co-stimulator of the antigen presentation by ARH5.5, while that of CD2/LFA-3 had only little stimulatory effect. These results suggest that the interaction between some of the adhesion molecules on APC and Th can cross the species barrier. The experimental system presented here is simple and useful for analyzing human APC function, Key words: Xenogeneic antigen presentation; Antigen-presenting cell function; Co-stimulatory molecule; (Human) *Corresponding author: Yasuko Tsunetsugu-Yokota, M.D., Ph.D., Department of Immunology, National Institute of Health, 1-23-1, Toyama, Shinjuku-ku, Tokyo 162, Japan. Tel.: 3-5285-1111 (ext. 2132); Fax: 3-5285-1150. Abbreviations: Th, helper T cell; APC, antigen-Presenting cell; MHC, major histocompatibility complex; OVA, ovalbumin; HEL, hen egg lysozyme; MMC, mitomycin-C. SSD1 0 1 6 5 - 2 4 7 8 ( 9 4 ) 0 0 0 2 5 - M

separately from T cell function, especially when the dysfunction of APC associated with viral infection with human tropism is considered.

2. Introduction Antigen recognition by T cells is the first crucial step of the immune response to pathogenic microbial agents. Usually in a virus infection, major histocompatibility complex (MHC) gene expression is up-regulated by cytokines such as interferons, which results in enhanced recognition of infected cells by responding T cells [1]. However, some viruses are known to have developed mechanisms to escape T-cell recognition by modulating MHC antigen [2,3] or cytokine expression [4]. Thus to understand the pathogenesis of viral infection, it is important to know the process of initiation of the immune response to virus infection. Analyzing the antigen-presenting cell (APC) function in humans by using primary T cells and monocyte/macrophages from the same individual may produce a problem because of the inconsistent results among individuals. On the other hand, the highly complex nature of human MHC compared to murine MHC makes it difficult to establish the appropriate combination of an HLA-matched APC line and a T-cell clone. Here we used a human B-cell clone expressing a murine MHC class II gene and analyzed its ability to present antigen to mouse T cells. Using this system, it is possible to see the effect of a humantropic virus on the antigen presentation, focusing on APC function by separating it from the T-cell activation process. For example, APCs are thought to play an important role in a variety of T-cell dysfunctions

associated with HIV infection [5]. In such a case, the readout system using mouse T cells has a great advantage because mouse T cells are not infected with HIV. The experimental system in this study may help us to clarify the molecular mechanism(s) of the APC function altered by infection with various pathogens or drugs.

3. Materials and Methods

3.1. Cell culture

ARH is an Epstein-Barr virus-positive human Bcell line [6,7] and is maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum and antibiotics. A mouse T-cell hybridoma, KE17-1, and a T-cell clone, AOIT.2.11, were obtained through the courtesy of Dr. K. Ogasawara (Hokkaido University, Sapporo, Japan) and Dr. A. Ametani (University of Tokyo, Tokyo, Japan), respectively. KE17-1 is specific for the immunogenic 16-mer peptide (AEGDSYTLANKRKGFT) of pigeon cytochrome C, and AOIT.2.11 recognizes the 23-mer peptide (NLCNIPCSALLSSDITASVNCAK) of hen egg lysozyme (HEL); both clones are restricted by I-A k. LK35.2, obtained through Dr. K. Ogasawara, is a mouse B-cell hybridoma which expresses both I-A k and I-E k. The OVA-21 T-cell clone was established from mouse inguinal lymph node T cells after immunization with ovalbumin (OVA) emulsified in complete Freund adjuvant (Difco Laboratories, MI). 3.2. Plasmid and reagents

The I-A k expression plasmid containing a neomycin-resistant gene was obtained from Dr. M. Watanabe (Mitsubishi-Yuka, Tokyo) [8]. OVA (fraction V), mitomycin-C (MMC), and chloroquine were purchased from Sigma (MO). Paraformaldehyde and G418 were purchased from Wako Junyaku (Tokyo). Anti-mouse LFA-1 (KBA), anti-mouse CD2 (RM21), anti-human ICAM-1 (HA58), and anti-human LFA-3 (TS 2/9) monoclonal antibodies were produced in Juntendo University, Tokyo. 3.3. Transfection

The I-A k expression plasmid was introduced into ARH by the electroporation using Gene Pulser (BioRad Laboratories, CA) under the condition of 960 /~F capacitance at 290 V. The cells were selected with 1 mg of G418 per ml and resistant cells were sub74

cloned by limiting dilution. To examine the expression of I-A k, cells were stained with biotinylated anti-I-Ak antibody (10-2.16) followed by FITC-avidin (Vector Laboratories, CA) and analyzed with a FACScan (Becton Dickinson, Fujisawa, Japan). 3.4. Assay for antigen-presenting function

APCs were treated with MMC (25 #g/ml) for 45 min at 37°C, followed by extensive washing and cultured with murine helper T-cell (Th) clone in the presence of antigens. Twenty-four hours later, the culture supernatant was collected and assayed for IL-2 content by using a mouse IL-2-dependent cell line, MTH (kindly provided by Dr. Y. Hitoshi, Institute of Medical Science, University of Tokyo). The proliferation of the cells in triplicate cultures was measured by the uptake of [3H]thymidine (ICN Pharmaceuticals, CA). The variation among those measurements is generally less than 5% of the mean.

4. Results

4.1. Expression of I-A k molecules on a human B-cell line

The plasmid-containing genomic I-A k • and fl chain was used to transfect ARH cells by electroporation and G418-resistant clones were obtained. After limiting dilution, one of the subclones, ARH5.5, which expresses a high level of I-A k, was selected. As shown in Fig. 1, the level of expression of I-A k molecules on ARH5.5 was comparable to that on LK35,2 cells, an I-Ak-positive mouse B-cell hybridoma. 4.2. Antigen presentation by ARH5.5 to murine Th cells

The function 'of peptide antigen presentation of MMC-treated ARH, ARH5.5 and LK35.2 was studied by using I-Ak-restricted, peptide-specific, murine Th cells (KE17-1 and AOIT.2.11), in the presence of antigenic peptides. The amount of IL-2 produced by activated KE17-1 or AOIT.2.11 was measured by :the [3H]thymidine uptake by MTH cells. The result of representative experiment is shown in Table 1. The supernatant of activated KE17-1 in the presence of ARH5.5 and LK35.2 induced proliferation of MTH tO the level of 35,089 and 52,885 cpm, respectively. Similar activation was observed when AOIT.2.11 was used. The supernatant of activated

1.1(35.2

/\i

ARH5.5

ARII

/:"

i

.......

Log Fluorescence Intensity Fig. 1. I-A k expression on LK35.2, A R H , and the I-Ak-transfectant, ARH5.5. Cells were reacted with a biotinylated-monoclonal antibody against I-A k (10-2.16) followed by FITC-avidin. Stained cells were analyzed with a FACScan: - - - , cells without antibodies; ........ , cells with second antibody only; - - - , cells with both antibodies.

AOIT.2.11 in the presence of ARH5.5 and LK35.2 induced proliferation of M T H to the level of 60,464 and 48,160 cpm, respectively. In both cases, the proliferation of M T H induced by the supernatant of activated Th cells in the presence of untransfected A R H was within the background level. Thus the human IA k transfectant, ARH5.5, was shown to be able to present a peptide antigen as well as the murine I-A kpositive LK35.2.

4.3. Processing of OVA antigen by ARH5.5 The presentation of OVA protein by ARH5.5 was assessed by using the OVA-21 T cell clone. As shown in Fig. 2, ARH5.5 was able to present OVA antigen to murine Th cells in a dose-dependent manner, although less efficiently (2- to 3-fold lower proliferation of MTH) than LK35.2.

TABLE 1 P R E S E N T A T I O N O F P E P T I D E A N T I G E N S BY ARH5.5 Th

Antigen

APC

Antigen concentration (#M)

In general, protein antigens need to be processed to generate peptides in the endosomal compartment so that they can associate with M H C molecules. Since ARH5.5 was able to present the native OVA antigen, such processing might be the mode of operation by the APC. Alternatively, ARH5.5 may just capture the naturally degenerated OVA antigenic fragments in the OVA preparation. In this case, the processing by ARH5.5 is not necessary for antigen presentation. Since ARH5.5 requires more OVA antigen than LK35.2 for comparable activation of murine Th cells as stated above, it is necessary to clarify the processing function of ARH5.5. For this purpose, chloroquine treatment, which is thought to inhibit proteolytic enzyme activity in the endosome environ~

5

"~

4

O O

.fi 0

3

8 2

30

3H-TdR incorporation by M T H (cpm) KEI7-1

cyt-C peptide*

AOIT.2.11

H E L peptide**

ARH ARH5.5 LK35.2 ARH ARH5.5 LK35.2

2623 1237 1321 2210 975 524

2117 35,089 52,885 5050 60,464 48,160

*Cytochrome C-derived peptide ( A E G D S Y T L A N K R K G F T ) . **HEL-derived peptide ( N L C N I P C S A L L S S D I T A S V N C A K ) .

ARH5.5

LK35.2

Fig. 2. Presentation of protein antigen by ARH5.5 and LK35.2. An OVA-21 murine Th clone (5 x 104 cells) was cultured with MMC-treated ARH5.5 or LK35.2 cells (5 × 104) in the presence or absence of OVA antigen. The culture supernatant was collected on the following day and the a m o u n t of IL-2 was measured by the proliferation of a murine 1L-2-dependent clone, M T H : m , none el, 2 mg/ml; m , 5 mg/ml.

75

[3H]-thymidine incorporation b y M T H (cpm) 102 103 104 105 I

I

i

I

I

, , 'l

t

,

,

i i i [ , I

.

.

.

.

.

.

.

% inhibition of T cell activation 0

I

OVA-21

÷mCD2

CQ 1.2raM

+hlCAM-1

CO 3.6mM

+hLFA-3

AOIT.2.11

40

60

80 I

100

+mLFA-1

CQ 0.4raM

PFA 0.1%

20

+mCD2+hLFA-3 +mLFA-I+hlCAM-1 +NMS

PFA 0.1%

Fig. 3. Inhibition of OVA antigen presentation by chloroquine and paraformaldehyde. MMC-treated ARH5.5 cells were treated with either chloroquine for 15 min or 0.1% paraformaldehyde for 30 min, washed extensively, then cultured with murine Th cells and respective antigens. OVA was used at 5 mg/ml, and HEL peptide at 10 #M. CQ, chloroquine; PFA, paraformaldehyde.

ment, and paraformaldehyde fixation of the membrane were carried out. By treatment with chloroquine for 15 min before addition of the OVA pulse, the APC function of ARH5.5 was inhibited in a dosedependent manner (Fig. 3). ARH5.5 was also treated with paraformaldehyde for 30 min, washed 3 times and cultured with OVA-21 or AOIT.2.11 in the presence of the respective antigen (OVA or H E L peptide). As shown in Fig. 3, while the activation of AOIT2.11 was not affected, that of OVA-21 was dramatically inhibited, demonstrating that ARH5.5 did process OVA antigen to present antigenic peptides to murine Th cells. The experiments were repeated 3 times and found to be reproducible.

Fig. 4. The effect of antibodies against adhesion molecules on APC function of ARH5.5. MMC-treated ARH5.5 cells, OVA-21 Th cells, and OVA antigen (5 mg/ml) were cultured in the presence of antibodies against adhesion molecules: m, 5 #g/ml, [] 25/lg/ml. The percent inhibition of T cell activation was calculated as follows: IL-2 production in the absence of antibody [1] x 100. IL-2 production in the presence of antibody

used for blocking the activation of T cells. A representative experiment is shown in Fig. 4. The antibody against LFA-1 on the murine Th and ICAM-1 on the human APC inhibited T-cell activation by 65% and 46%, respectively. The combination of these antibodies had an additive blocking effect. On the other hand, the blocking effect of the antibody against CD2 on the murine Th was marginal and the antibody against LFA-3 on the human APC did not block the activation of T cells. However, the combination of these antibodies showed 53% inhibition. These results indicate that some adhesion molecules cross-interact and are able to act as co-stimulatory factors despite the xenogenic combination.

4.4. Involvement o f adhesion molecules in the interaction between human A P C and murine Th 5. Discussion

Besides T-cell receptor-MHC interaction, several adhesion molecules on the surface of Th and APC are known to act as co-stimulatory factors for T-cell activation. Also in this system, it is possible that adhesion molecules of different species can interact and then deliver the appropriate signal for T-cell activation. In order to study which molecules are involved in this co-stimulatory interaction between human and mouse, antibodies against adhesion molecules were 76

We have established an I-Ak-transfected human Bcell clone, ARH5.5, and demonstrated that this cell line is able to process an immunogenic antigen and present it to murine Th cells. Although we observed a similar ability of ARH5.5 to present peptides to murine APC (LK35.2), a much higher concentration of OVA antigen was required for ARH5.5 to activate an OVA-21 murine Th clone than that required for

the murine APC. Since A R H 5 . 5 expresses high levels o f H L A antigens (data not shown), these h u m a n and mouse M H C class II molecules m a y compete for the binding of the processed O V A peptides in the endosome. Alternatively, because proteolytic enzymes have different specificities in mice and humans, O V A peptides processed by A R H 5 . 5 m a y not completely fit the binding pocket o f the I-A k molecule, resulting in the generation o f low affinity O V A peptides and inefficient presentation by A R H 5 . 5 . In general, T-cell responses are k n o w n to be more efficient within a species than across species [9,10]. Studies using murine cells transfected with h u m a n M H C molecules revealed inefficiency in recognition by h u m a n T cells [11,12]. As one possibility, the interaction between h u m a n adhesion receptors on the T cell and the murine counter-receptors on the targets is poor. This was supported by the fact that the efficiency o f M H C class-II-transfected murine L cells as A P C was m u c h improved by transfecting h u m a n I C A M - 1 [13]. To determine the importance o f adhesion molecules in the xenogeneic cell interactions described here, the blocking effect by antibodies against adhesion molecules was studied. J o h n s t o n et al. [14] f o u n d that h u m a n I C A M - 1 did not bind to the murine LFA-1 by studies using purified h u m a n I C A M - 1 on solid substrates. In contrast, we observed that Tcell activation was blocked by antibodies against murine LFA-1 or h u m a n I C A M - 1 . These results suggest that in close contact o f T-cell receptor and M H C molecules, weak interaction o f murine LFA-1 and h u m a n I C A M - 1 could occur and this m a y be sufficient for T cells to be activated. Since murine LFA-1 is k n o w n to interact with other relevant molecules such as I C A M - 2 or -3 [15], and h u m a n I C A M - 1 with C D 4 3 [16], these alternative ligands m a y also be effective for the T-cell activation. The interaction between C D 2 and L F A - 3 is also important as a costimulatory factor [17]. In this study, the antibody against murine C D 2 or that against h u m a n L F A - 3 had marginal or little blocking effect, but in combination they blocked T-cell activation significantly. Since the main ligand for murine C D 2 is reported to be CD48 [18], that m a y explain why anti-human L F A - 3 had no effect but anti-murine C D 2 a n t i b o d y had some. P r o b a b l y murine C D 2 can interact weakly with h u m a n CD48. However, the mechanism o f the substantial inhibition observed with the combination o f these antibodies is not clear and other adhesion

molecules m a y also play a role on the T-cell activation. This xenogenic A P C system will be applied for the evaluation o f the effect o f i m m u n o - m o d u l a t o r y drugs or h u m a n tropic virus on A P C function in human.

Acknowledgements We would like to thank Dr. A. Meyerhans for discussion. This w o r k was supported by a grant from the Ministry o f Public Health and Welfare, Japan.

References [1] Maudsley, D.J. and Pound J.D. (1991) Immunol. Today 12, 429. [2] Browne, H., Smith, G., Beck, S. and Minson, T. (1990) Nature 347, 770. [3] Burgert, H.-G. and Kvist, S. (1985) Cell 41,987. [4] Gooding, L.R. (1992) Cell 7l, 5. [5] Meyaard, L., Shuitemaker, H. and Miedema, F. (1993) Immunol. Today 14, 161. [6] Yano, A., Aosai, F., Ohta, M., Hasekura, H., Sugane, K. and Hayashi, S. Parasitology 75, 311. [7] Yano, A., Ohno, S., Norose, K., Baba, T., Yamashita, K., Aosai, F. and Segawa, K. (1992) Int. Arch. Allergy Immunol. 98, 13. [8] Watanabe, M., Shinohara, N., Ochi, A. and Hozumi, N. (1986) Immunol. Lett. 13, 237. [9] Asantila, T., Vahala, J. and Toivanen, P. (1974) lmmunogenetics 1,272. [10] Barbosa, J.A., Mentyer, S.J., Minowada, G., Strominger, J.L., Burakoff, S.J. and Biro, P.A. (1984) Proc. Natl. Acad. Sci. USA 81, 7549. [11] Van de Rijn, M., Bernabeu, C., Royer-Pokora, B., Weiss, J., Seidman, J.G., de Vries, J., Spits, H. and Terhorst, C. (1984) Science 226, 1083. [12] Bernabeu, C., Maziarz, R., Spits, H., de Vries, J., Burakoff, S.J. and Terhorst, C. (1984) J. Immunol. 133, 3188. [13] Altmann, D.M., Hogg, N., Trowsdale, J. and Wilkinson, D. (1989) Nature 338, 512. [14] Johnston, S.C., Dustin, M.L, Hibbs, M.L. and Springer, T.A. (1990) J. Immunol. 145, 1181. [15] De Fougerolles, A.R. and Springer T.A. (1992) J. Exp. Med. 175, 185. [16] Rosenstein, Y., Park, J.K., Hahn, W.C., Rosen, F.S., Bierer, B.E. and Burakoff, S.J. (1991) Nature 354, 233. [17] Bierer, B.E., Peterson, A., Barbosa, J., Seed, B. and Burakoff, S.J. (1988) Proc. Natl. Acad. Sci. USA 85, 1194. [18] Kato, K., Koyanagi, M., Okada, H., Takahashi, T., Williams, A.F., Okumura, K. and Yagita, H. (1992) J. Exp. Med. 176, 1241.

77