Effector functions of CD8-positive guinea pig T ...

Cellular Immunology 222 (2003) 134–143 www.elsevier.com/locate/ycimm

Effector functions of CD8-positive guinea pig T lymphocytes Hubert Sch€ afer,a,* Rik J. Scheper,b Bettina Borsorf,a and Reinhard Burgera a

Department of Immunology, Robert Koch Institute, Nordufer 20, 13353 Berlin, Germany b Department of Pathology, Free University, Amsterdam, The Netherlands Received 22 January 2003; accepted 17 April 2003

Abstract The response of guinea pig T lymphocytes to different stimuli was analysed with focus on the functions of CD8-positive T cells, which so far had been poorly defined in this animal model. For identification and purification of guinea pig cytotoxic T lymphocytes, three monoclonal antibodies, directed against the CD8 differentiation antigen were characterized and compared with respect to expression pattern and biochemical characteristics of the corresponding cell surface antigen. The antibodies were used for the identification of the cytotoxic T lymphocyte subpopulation within alloreactive T cell lines, and for the depletion of CD8-positive cells in in vitro assays. Purified CD4- and CD8-positive cells were tested for their ability to proliferate in response to antigen, mitogen or anti-guinea pig Thy-1 monoclonal antibodies. Both, CD4- and CD8-positive cells showed IL-2 release and subsequent proliferation after polyclonal stimulation. Cytotoxic activity in CD8-positive alloreactive T cells was expressed in vitro only after repeated stimulation. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: Guinea pig; Effector function; CD8; Cytotoxic; T lymphocytes; Proliferation; IL2; CFSE

1. Introduction To understand the mechanism of immune reactions, knowledge of the cell populations involved, and of their effector functions, is needed. Due to their different effector functions, CD4- and CD8-positive T lymphocytes show marked differences in their response after antigen- or mitogen-induced activation (reviewed in [1]). CD4-positive T cells express surface molecules needed to provide B cell help and, upon activation, these ‘‘helper’’ or ‘‘inducer’’ T cells upregulate cytokine secretion in a pattern according to their TH1 - or TH2 phenotype [2]. CD8-positive T cells in contrast, are characterized by the expression of lytic molecules such as granzymes and perforin, and differentiate into cytotoxic effector cells. Cytotoxic function however, is not fully restricted to CD8-positive T cells, and the release of regulatory cytokines is not an exclusive function of CD4-positive cells. It has been shown for example, that CD4-positive T cells can exert cytotoxic * Corresponding author. Fax: +49-30-4547-2602. E-mail address: [email protected] (H. Sch€afer).

effector functions and express most of the molecules associated with the cytolytic activity of CD8-positive cells [3]. CD8-positive T cell clones on the other hand, are able to release regulatory cytokines such as IL2, IL4, and IFN-c [4,5]. In general, there are still marked differences in the ability of these cell populations to express effector molecules, secrete cytokines, and to interact with other cells. In cloned mouse spleen cells a clear segregation of cytolytic activity and IL2-producing activity was observed [6], CD8-positive T cell clones frequently showed cytolytic activity and rarely produced IL2, whereas the opposite was found for CD8-negative clones. In human T cells both CD4- and CD8-positive subpopulations were able to produce IL2, but differential stimulation requirements for both subsets were identified [7]. Functional differences in T cell subpopulations are observed on at least three levels: (1) individual subsets of T cells show different abilities to carry out effector functions and to release cytokines, (2) the induction of specific functions in individual subsets depends on the mode of stimulation, and (3) the functional characteristics of individual T cell subsets are different between species.

0008-8749/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0008-8749(03)00112-6

H. Sch€afer et al. / Cellular Immunology 222 (2003) 134–143

The present study was carried out to analyse the functional capabilities of CD4- and CD8-positive T cells in the guinea pig. The guinea pig provides an important animal model for numerous infectious diseases. Insight into T lymphocyte effector functions on the cellular level might help to explain some peculiarities of the guinea pig immune response such as the high susceptibility to intracellular pathogens, including Mycobacterium tuberculosis [8].

135

2.4. Peritoneal exudate cells Twenty-five milliliters of a sterile solution of soluble starch (2% in 0.9% NaCl) or 15 ml of sterile mineral oil was injected into the peritoneum of adult guinea pigs. After 3–5 days, the animals were sacrificed and the peritoneal cells were washed out with 150 ml of ice-cold HBSS or RPMI. 2.5. Purification, biotinylation, and FITC-labelling of monoclonal antibodies

2. Materials and methods 2.1. Animals Guinea pigs of the inbred strain 2, strain 13, and JY-1 were bred in our animal facility. Animals from inbred strain Rockefeller were kindly provided by Hoffmann La Roche, Basel, Switzerland. The MHC antigens of these strains are listed in Table 1 (compiled from several sources [9–11]).

B607 hybridoma cells were grown to the late log phase in spinner cultures and IgG (mouse IgG1) was purified from the supernatant using protein G coupled to sepharose beads (Amersham Pharmacia Biotech, Freiburg, Germany). CT6 was purified from ascites by protein G affinity chromatography. MSgp6 (mouse IgM) was purified from supernatant using ‘‘HiTrap IgM’’ purification columns (Amersham Pharmacia Biotech, Freiburg) according to manufacturerÕs instructions.

2.2. Monoclonal antibodies 2.6. Fluorescence staining and flow cytometry The generation and a preliminary characterization using immunohistological staining of CT6 and MSgp6 has already been described [12,13]. The mAb H155 [14] identifies the guinea pig CD4-antigen, H159 reacts with mature T cells [15], H154 is directed against the guinea pig Thy-1 antigen (CD90) [16]. Mab B607 was obtained after immunizing female BALB/c mice with guinea pig lymphocytes and fusion of immune spleen cells with the P3X63Ag8.653 myeloma line following standard procedures. The isotype of the antibodies was determined using a mouse monoclonal isotyping kit (Calbiochem, San Diego, CA, USA). 2.3. Lymph node lymphocytes Lymph node cells from immunized guinea pigs (50 lg OVA in 50 ll of PBS, emulsified with 50 ll of complete FreundÕs adjuvant, CFA) were incubated in a nylon wool column for 45 min at 37 °C. T cells were eluted by rinsing the column with prewarmed RPMI. The obtained cell population contained more than 90% T lymphocytes. Table 1 MHC class I and class II antigens in the guinea pig Strain

2 13 Rockefeller JY-1

MHC-region D/K (class I)

Ia (class II)

B.1 B.1 B.3 B.3

2, 1, 1, 1,

S.1 S.1 – –

4, 3, 3, 6,

5, 5, 5, 7,

6 6, 7 6, 7 8

Cells were adjusted to 2–5 106 /ml in PBS containing 2% fetal calf serum and 0.01% sodium azide. Fifty microliters of the cell suspension was mixed with 50 ll of antibody and incubated on ice for 30 min. After washing twice, 50 ll of a 1:50 dilution of FITC-labelled rabbitanti-rat (or anti-mouse) Ig F(ab)2 , (Jackson, USA) were added to each sample. Cells were incubated for 30 min on ice, washed three times, and resuspended in 200 ll of PBS. For two-colour immunofluorescence anti-rat phycoerythrin was used in the first step, followed by a second labelling with FITC-labelled mAb or with biotinylated mAb and FITC-labelled streptavidin. Fluorocytometric analysis was carried out using a ‘‘FACScalibur’’ cytometer (Becton–Dickinson, USA). 2.7. Cell surface labelling, immunoprecipitation, and autoradiography A mixture of lymph node lymphocytes and thymocytes were labelled with 125 I by the lactoperoxidase method as described previously [14]. The lysates were precleared twice by incubation with 10% of ‘‘Pansorbin’’ (suspension of fixed Staphylococcus aureus bacteria strain Cowan I, Calbiochem) and rabbit-anti-mouse Igcoated ‘‘Pansorbin’’ for 45 min on ice each. Immunoprecipitation was performed by precoating 100 ll of suspended S. aureus bacteria in precipitation buffer (10% v/v in PBS containing 1% Triton X-100, 0.5% sodium desoxycholate, and 0.5% BSA) with polyclonal rabbitanti-mouse Ig (Dianova, Hamburg, Germany). In parallel, 100 ll of precleared iodinated T cell lysate were

136


incubated with 2–5 ll of monoclonal antibody (ascites) for 1 h under continuous agitation at 4 °C. Immune complexes were isolated by adding 100 ll of rabbit-antimouse Ig precoated S. aureus suspension (as above). After keeping the suspension for 45 min on ice, bacterial cells were pelleted by centrifugation (12,000g in an Eppendorf centrifuge), washed two times in precipitation buffer and three times in washing buffer (identical to precipitation buffer, but without BSA). The pellet was resuspended in 25 ll of SDS sample buffer, boiled for 5 min, centrifuged for 5 min at 12,000g, and the supernatant was separated on an SDS gradient gel (8–15% acrylamide). Gels were dried and autoradiographed on X-ray films (Kodak X-Omat, AR) for 2–14 days. 2.8. Alloreactive T cell lines All cell cultures were carried out using RPMI medium supplemented with 10% fetal calf serum, glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 lg/ml, all supplements obtained from Gibco BRL, Karlsruhe, Germany), and 2-mercaptoethanol (5 105 M). Alloreactive T cell lines were established by setting up mixed lymphocyte cultures of nylon wool- or rayon wool-purified lymph node lymphocytes (LNL) (3 106 ) stimulated by allogeneic oil induced, irradiated peritoneal exudate cells (PEC) (1 106 ) in 24-well plates. After 5–7 days of culture, proliferating T cells were harvested, adjusted to 1 106 /ml, seeded into 24-well plates (1 ml), and restimulated by addition of 1 ml irradiated PEC (1 106 /ml). Exogenous recombinant IL2 (10 U/ml) of human (Amersham, Pharmacia) or guinea pig origin (supernatant from gp IL2-transfected cell lines) was provided upon restimulation, but not in the primary MLR, as an external growth factor. For long-term cultivation, T cells were restimulated weekly following the same protocol. 2.9. T cell proliferation assay Nylon wool purified T lymphocytes were adjusted to 3 106 cells/ml in RPMI medium (supplements as above) and 100 ll/well were distributed into 96-well plates. Cells were activated by addition of ConA (10 lg/ ml), PHA (5 lg/ml) or anti-Thy-1 monoclonal antibody H154 (crosslinked by anti-rat Ig and costimulated with 5 108 M PMA). At day four of the culture [3 H]thymidine was added and cells were harvested 24 h later for measurement of incorporated radioactive thymidine by liquid scintillation counting. 2.10. Measurement of T cell proliferation using intracellular fluorescent labels and flow cytometry [17] A stock solution of 5 mM 5-(and 6)-carboxy fluorescein diacetate, succinimidyl ester (CFDA-SE, Sigma)

was prepared in DMSO. One microliter of 5 mM CFDA-SE was added to 10 106 cells in 1 ml PBS. The sample was shortly mixed (vortex) and incubated for 2 min at room temperature. Ten milliliters of RPMI medium containing 10% FCS was added to stop the reaction and cells were washed twice in RPMI with 10% FCS. Labelled cells were adjusted to 3 106 /ml, seeded into 24-well plates at 1 ml/well, and 1 ml of culture medium containing the appropriate amount of mitogen was added. Cells were recovered 3–4 days after stimulation, stained for expression of differentiation antigens as described, and analysed using a FACScalibur cytometer (Becton–Dickinson). 2.11. In vitro assay for cellular cytotoxicity T lymphocytes from alloreactive cell lines were harvested, washed in RPMI, and resuspended at 3 106 /ml in culture medium. Target cells (106 peritoneal exudate cells in 1 ml of culture medium) were labelled with 100 lCi of 51 Cr (as sodium chromate, NEN, Dreieich, Germany) for 1 h at 37 °C. Labelled cells were washed, and resuspended at 1.5–6 105 cells/ml. Effector cells (100 ll) were plated in 96-well round bottom plates and 100 ll of the labelled target cells were added to the culture. Plates were incubated for 4 h at 37 °C in a humidified incubator, centrifuged for 5 min at 200g, and 100 ll of the supernatant was used to determine released radioactivity by gamma-scintillation counting (Auto Cobra Gamma Counter, Canberra Packard, Frankfurt, Germany). Spontaneous release was determined by incubating target cells with medium alone. Maximum lysis was quantitated by lysing target cells in 1% Triton X100. Specific cytotoxicity was calculated as follows: specific lysis ½% ¼

released activity ½cpm spont: rel: ½cpm 100: maximum lysis ½cpm spont: rel: ½cpm

2.12. Purification of T cell subpopulations by magnetic depletion T cells from spleens or lymph nodes were adjusted to 1 107 cells/ml in RPMI containing 10% FCS, and the proportion of CD4þ and CD8þ T cells in each suspension was estimated based on the fluorocytometric analysis. After preincubation with mAb H155 (for depletion of CD4-positive cells) or B607 (for depletion of CD8positive cells) cells were washed and a sufficient amount of antibody coated (anti-rat Ig or anti-mouse Ig, respectively) magnetic beads (Dynabeads, Dynal, Hamburg, Germany) was added to establish a ratio of at least four beads per cell to be depleted. Cells were incubated and separated according to manufacturerÕs instructions. The outcome of each depletion was examined by FACS analysis. As shown previously magnetic sorting resulted in 97–99% depletion of targeted cells [18].


3. Results 3.1. Flow cytometric analysis of anti-CD8 mAb Since the monoclonal antibodies against guinea pig CD8 had not been characterized in detail, an analysis of the reaction pattern and of the recognized antigen was carried out. LNL and thymocytes were stained with mAb B607, CT6, and MSgp6 as the primary, and with FITC-labelled rabbit-anti-mouse Ig as secondary antibody. All mAb reacted with about 25% of lymph node cells (Fig. 1A) showing that the antigen is expressed only on a subpopulation of mature T cells. In the thymus, in contrast, the majority (70–80%) of the cells were stained by these antibodies. Cumulative staining of lymph node cells with all three monoclonals did not result in an increased portion of antibody-positive T cells (Fig. 1A), indicating, that all antibodies reacted with the identical subpopulation of T cells. It was evident in all experiments, that B607 and CT6 stained more intensively than Msgp6, if comparable concentrations of antibody were used. Crossblocking experiments using biotinylated antibodies showed, that all antibodies recognized closely related structures on the cell surface. Both B607 and CT6 blocked the binding of all three mAb almost to background levels (Fig. 1B). MSgp6 in contrast, inhib-

137

ited the binding of the biotinylated antibodies CT6 and B607 only partially. Blocking studies were performed on lymph node cells using B607 and MSgp6, and on thymocytes using CT6 as blocked antibodies. 3.2. Two colour immunofluorescence Staining of LNL with a combination of monoclonal antibody H155 (directed against the guinea pig CD4antigen), and any of the putative anti-CD8 antibodies, resulted in an additional distinct fluorescence peak when compared to cells stained with H155 or with B607, CT6 or MSgp6 alone (not shown). Obviously, two non-overlapping populations were recognized among peripheral T cells: the number of positive cells in the combined staining was identical to the combined percentages of H155- and B607-positive cells in individual staining experiments. For a more detailed analysis of the cellular distribution of these cell surface markers two-colour immunofluorescence was performed. LNL and thymocytes were stained with the B607, CT6 or MSgp6 in the first step, and rabbit-antirat phycoerythrin as secondary antibody, followed by FITC-labelled anti-CD4. Mature T cells from lymph nodes were composed of two defined subpopulations expressing either CD4 or the B607-antigen (Fig. 2A). Some double-negative, but only very few double-positive cells were detected. In contrast, most of the

Fig. 1. (A) Cytofluorometric analysis of LNL. Nylon wool purified cells were stained with monoclonal antibodies in the first and FITC-labelled rabbit-anti-mouse Ig in the second step. All mAb stained a similar portion of mature T cells. Cumulative staining with all antibodies did not increase the proportion of positive cells, indicating, that all mAb reacted with the same cell population. (B) Crossblocking of monoclonal antibodies. Cells were incubated with purified mAb B607, MSgp6, CT6 or medium for 45 min. Directly labelled antibody was added and incubation was continued for another 45 min. Cells were analysed for binding of the directly labelled antibody. Crossblocking experiments were carried out on mature lymphocytes (represented for B607 and MSgp6) and on thymocytes (shown for CT6).

138


thymocytes, expressed both CD4 and the antigen identified by B607 or CT6 (about 65%, Fig. 2B). It should be noted, that in the same experiment double staining with MSgp6 and H155 produced less double positive thymocytes, but an increased number of CD4 single positive T cells were detected (not shown). After double staining with a T cell marker, almost all of the B607-positive lymph node T cells from immunized animals strongly co-expressed the antigen detected by H159 (mainly expressed on mature T cells), indicating that the vast majority of B607-positive cells were in fact T cells (Fig. 2C). The larger population of H159-positive, B607-negative cells obviously represents the T helper subset. 3.3. Biochemical characterization of the recognized antigen Thymocytes and LNL were surface-iodinated by the lactoperoxidase method and solubilized by detergent lysis. The lysates were preincubated with irrelevant an-

tibody, with mAb CT6 and, as a positive control, with MSgp4 (anti-MHC class I). Immune complexes were precipitated and analysed by SDS–PAGE (Fig. 3). CT6 precipitated a protein with a molecular weight of about 33 kDa when separated under reducing conditions (lane B, left panel). No distinct band was identified when the same precipitate was analysed under non-reducing conditions (lane B, right panel). The anti-MHC class I mAb (lane C) precipitated two proteins both under reducing and non-reducing conditions: the 45 kDa bands most likely representing the MHC class I protein, the band close to the 14 kDa marker represents the associated b-2 microglobulin. The expected CD8 dimers reported on human and mouse T cells were not identified or had co-migrated with the non-specifically precipitated material at about 60 and 100 kDa, which is also present in the control—lane (A). In a different experiment using lysates of biotinylated cells CT6 and B607 precipitated identical molecules of 33 kDa under reducing conditions (not shown), but again no evidence for additional proteins was found.

Fig. 2. Two colour immunofluorescence of LNL and thymocytes. (A) Lymph node cells were stained with different putative anti-CD8 mAb and antimouse PE in the first and with directly labelled anti-CD4 (mAb H155) in the second step. (B) Double staining of thymocytes with B607 and anti-CD4 (as described for A). (C) Double staining of lymphocytes with B607 and H159 (surface marker for T cells).


139

Fig. 3. Immunoprecipitation of guinea pig CD8 and MHC class I antigens. T lymphocytes were surface-labelled with 125 I using the lactoperoxidase method. Specific antigens were precipitated using control antibody (lane A), mAb CT6 (lane B), and MSgp4 (lane C) coated to fixed S. aureus bacteria. Precipitates were separated under reducing and non-reducing conditions in a PAA-gel. Co-precipitation of b2 microglobulin by the anti-class I mAb is indicated.

3.4. Phenotypic and functional characterization of longterm cultivated alloreactive T cell lines We had experienced in previous assays, that guinea pig T cells, stimulated in a primary 7 days MLC, did not express cellular cytotoxicity against the stimulating allogeneic strain. Only after repeated in vitro restimulation a cytotoxic response was observed. To set up cytotoxic T cell lines, T lymphocytes from different guinea pig inbred strains were stimulated by MLC with irradiated allogeneic PEC and kept in culture for several weeks by repeated restimulation and addition of recombinant human IL2. The reactive cells were phenotyped with respect to expression of CD4- or CD8antigens, the reactivity pattern was analysed by cytotoxicity and proliferation assays. As expected, the highest specific cytotoxicity was observed in a combination of strains differing in MHC class I antigens only, i.e., Rockefeller anti-strain 13 (Fig. 4) or vice versa. In other combinations, such as JY-1 anti-strain 13 (Table 2), CD8-positive cells rapidly disappeared after restimulation, although both MHC class I- and class II-restricted responses should be stimulated. The cytolytic activity of these cell lines decreased in parallel with the loss of CD8-positive cells. Alloreactive T cell lines were restimulated in weekly intervals by addition of allogeneic stimulator cell and recombinant human IL2. After cultivation for 17 days (2 times restimulated) and 40 days (5 times restimulated) cells were tested for cytotoxic activity and CD4/CD8phenotype.

Fig. 4. Phenotype and function of alloreactive T cell lines. T cell lines specific for MHC class I alloantigens or for MHC class I and class II alloantigens were initiated and cultivated for at least 4 weeks. (A) The phenotype of the responding T cells was determined by flow cytometry using mAb H159 (anti-T cell), H155 (anti-CD4), and MSgp6 (antiCD8). (B) Specific cytotoxicity was measured at a 15:1 effector:target ratio using 51 Cr-labelled macrophages as target cells. (C) Proliferation was quantified measuring the uptake of tritiated thymidine 3 days after restimulation over a 24 h period.

3.5. Induction of proliferation and IL2-release in purified T cell subpopulations The proliferation of activated T lymphocytes strongly depends on the presence of growth factors such as IL2 or IL4. For most T cells, these factors are supplied in an autocrine manner, i.e., the same cells that depend on IL2 for proliferation release this cytokine early after activation. Not all T cells however produce sufficient amounts of cytokines to induce and sustain proliferation, but depend on exogenous sources. The ability of semi-purified (by magnetic depletion) T cell subsets to respond to different mitogenic stimuli was tested by stimulation with the mitogens ConA and PHA, and with the mitogenic monoclonal anti-Thy-1 (H154) antibody

140


Table 2 Cytotoxic activity and phenotype of alloreactive long-term T cell lines T cell line

Cultivation (days)

Cytotoxicitya (% specific lysis)

Phenotypeb (% of total) H159þ

H155þ

Msgp6þ

Rock. anti-13

17 40

47 55

84 80

52 32

33 44

JY-1 anti-13

17 40

38 8

89 97

60 96

30 0.1

a b

Determined by 51 Cr-release against strain 13 PEC, E : T ¼ 15 : 1, SD