Immunobiol., vol. 184, pp. 359-371 (1992)
1 Department of Medical and Physiological Chemistry, University of Uppsala and 2Department of Inflammation Research, Kabi Pharmacia, Uppsala, Sweden
Activated Type II Collagen Reactive T Cells are not Eliminated by in vivo Anti-CD4 Treatment. Implications for Therapeutic Approaches on Autoimmune Arthritis TOM]' GOLDSCHMIDT 1,2, MIKAELANDERSSON 1, VIVIANNEMALMSTROM 1, and RIKARD HOLMDAHL 1 Received October 14, 1991 . Accepted November 18, 1991
Abstract Activation of CD4+ T cells plays an important role in type II collagen (CII) induced arthritis (CIA). The CD4+ T cell dependency is demonstrated by anti-CD4 antibody treatment which suppresses CIA in mice if injected before CII immunization. The same anti-CD4 treatment at a later stage does not suppress CIA, despite extensive elimination of peripheral CD4+ T cells. A possible explanation for this discrepancy is that activated T cells might not be as easily influenced by the anti-CD4 antibodies as resting T cells. To address this question, the proliferative capacity of CII reactive CD4+ lymph node (LN) T cells, in mice treated with antiCD4 antibodies before or after the CII immunization, was analyzed. In mice treated before immunization the capacity of LN cells to proliferate in vitro was markedly suppressed while in mice receiving anti-CD4 treatment after immunization it was retained. Flow cytometric analysis revealed that the anti-CD4 treatment before and after immunization reduced the number of CD4+ LN T cells to the same level. The small population of CD4+ LN cells which were left after anti-CD4 treatment of naive mice all expressed CD44, a marker for previously activated T cells in mice. We propose that activation render CII reactive T cells more resistant to anti-CD4 treatment than virgin T cells are and suggest that the lack of therapeutic effect of late anti-CD4 treatment in CIA does not necessarily implicate that CD4+ T cells are unimportant in that stage of the disease.
Introduction Type II collagen-induced arthritis (CIA) is an experimental tissuespecific autoimmune disease originally characterized in mice and rats which is considered to be an important model for rheumatoid arthritis (RA) (1, 2). CIA is induced by immunization with native type II collagen (CII), which in mice cause arthritis in peripheral joints within three to eight weeks. The pathogenic mechanism involves activation of CD4+ T cells as implicated by the association of arthritis susceptibility to certain MHC class II haplotypes Abbreviations: ClI = type II collagen; CIA = collagen-induced arthritis; RA = rheumatoid arthritis; Ab = antibody; mAb = monoclonal antibody; LN = lymph node; PCA = Freund's complete adjuvant; MHC = major histocompatibility complex
360 . T.
J. GOLDSCHMIDT, M. ANDERSSON, V. MALMSTROM, and R. HOLMDAHL
(3, 4) and the appearance of CD4+ T cells in early stages of arthritis (5). Moreover, the development of CIA can be prevented by injection of antiCD4 antibodies at the time of immunization (6, 7). However, it has been more difficult to affect the disease with anti-CD4 treatment at a later point of time (6,7), when the immune response to cn, shown by the presence of activated T cells and anti-Cn autoantibodies, has already been triggered (8,9). The reason why not anti-CD4 treatment at later stages have been therapeutically effective is unknown. It may be that CD4+ T cells only play a role in the induction of B cells to produce pathogenic autoantibodies. However, mAb treatment to CD4 usually does not result in complete elimination of CD4+ cells (10, 11). An alternative explanation for the absence of therapeutic effects in CIA may be that the monoclonal antibody treatment less efficiently affects activated T cells. We investigated if the occurrence of cn reactive T cells differs in mice which had been anti-CD4 treated before or after immunization with rat CII. The presence of cn reactive T cells was measured by an in vitro proliferation assay and the elimination of CD4 + T cells was analyzed by flow cytometry. The experiments were carried out in CIA susceptible DBA/1 mice in which anti-Cn antibody response and arthritis development were evaluated. Moreover, we studied whether T cells resistant to anti-CD4 or anti-CD8 mAb treatment possessed an activated phenotype by analyzing T cells, left after treatment of unimmunized mice, for the expression of CD44, a cell surface molecule expressed at high density on activated memory T cells (12).
Materials and Methods Induction and evaluation of CIA DBA/I] mice, originally obtained from] ackson Laboratories Inc, were bred and kept at the Biomedical Center in Uppsala and DBA/lOla, derived from Harlan Olac Ltd (Blackthorn, UK), were bred and kept at the Department of Pathology in Uppsala. In all experiments were used 8-12 weeks old, age matched, male mice. Adult thymectomy was done under pentobarbital anesthesia (Mebumal, ACO Likemedel, Goteborg, Sweden) and the mice were allowed to recover two weeks after surgery before the experiments started. Rat cn was prepared by pepsin digestion of a rat chondrosarcoma and subsequent purification with selective salt precipitation and ion exchange chromatography on DEAE-Sepharose (13). For immunization native rat cn was dissolved in 0.1 M acetic acid +4 DC and emulsified 1:1 on ice with Freund's complete adjuvant (FCA) (Difco, Detroit, MI, USA) to a final concentration of 1 mg/m!. Mice that were used in the arthritis experiment were injected intradermally at the base of the tail with 50 [11 containing 50 [1g of rat cn emulsified 1:1 in FCA. Mice that were used for the proliferation and flow cytometric analyses were immunized intradermally at the base of the tail and both hind foot pads within all 150 [1g rat cn. Arthritis development in the four paws was judged by a macroscopic scoring system ranging from 0 to 3 (1 = swelling and/or redness of one toe or finger joint, 2 = two or more joints involved and 3 = severe arthritis in the entire paw).
Activated T Cells not Eliminated by Anti-CD4 Treatment . 361
Quantification of anti-CIl antibodies in serum Sera were collected individually and stored at -20 DC until assayed. For the quantification of anti-Cn reactive autoantibodies in serum, a modified standard ELISA technique was used (4). In brief, micro-ELISA plates (Dynatech, Plochingen, Germany) were coated with 10 [Lg/mlof native rat cn. All tests were carried out in duplicates, and the standard deviations did not exceed 10 %. The amount of bound antibody was estimated after incubation with a goat antimouse IgG (H + L) affinity purified antibody conjugated to alkaline phosphatase 0ackson Immunoresearch Laboratories, West Grove, PA, USA). The subsequent quantification of bound enzyme was performed with a paranitrophenol containing substrate buffer in a Titertek Multiscan® spectrophotometer. To estimate the amount of anti-Cn reactive antibodies present in the serum samples, an affinity purified DBA/1 anti-Cn antibody preparation was used as standard and titrated in parallel with the unknown serum samples.
Monoclonal antibodies The following hybridomas were cultured; H129.19 (H129), which produce a rat IgG2a antibody specific for mouse CD4 (14); GKl.5, which produce a rat IgG2b which binds to mouse CD4 in a competitive fashion to H129 (15); MAR18:5, which produce a mouse IgG2a specific for the rat kappa chain (16); IM7.8.1, which produce a rat IgG2b to mouse CD44, was kindly provided by Dr. ROBSON MCDONALD, Lausanne (17). H129 hybridoma supernatant was concentrated and partly purified by ammonium sulphate precipitation between 25 and 45 percentage saturation, dialyzed to PBS, sterile filtered and stored at -20 DC until use. The specific H129 content was estimated by ELISA with an affinity purified H129 preparation as standard (4). GK1.5 was chromatographically purified (18). MARl.8 was affinity purified on Protein A-Sepharose (Pharmacia AB, Uppsala, Sweden), dialyzed with PBS and sterile filtered. The protein content was quantified by absorbance measurement at 280 nm.
Anti-CD4 antibody treatment in vivo The anti-CD4 treatment was performed by injecting 100 [Lg H129 i.p., which one hour later was followed by 200 [Lg MAR18.5 i.p. In this sandwich method the anti-rat kappa antibody MAR18.5 binds to H129 in vivo and enhances the elimination of CD4+ cells as earlier described (11). In the arthritis and proliferation experiments the anti-CD4 treatment was given either 10 days before or 5 days after immunization of the previously thymectomized mice. Before the immunization, the pre-treated mice were tested for the presence of persisting antiCD4 (H129) rat antibodies in serum. Micro-ELISA plates (Dynatech) were coated with goat anti-rat Ig (H + L) which had been mouse serum absorbed 0ackson Immunoresearch Laboratories, Inc.) and the serum titrated as described above. Bound rat monoclonal antibodies were subsequently detected by alkaline phosphatase conjugated, mouse serum absorbed, goat anti-rat Ig (H + L) 0ackson Immunoresearch Laboratories, Inc.). Purified H129 was used as standard and the calculations were performed as described above for anticn antibodies. In the experiment where the expression of CD44 on LN cells was measured, naive DBA/1] mice were anti-CD4 treated 8 days prior to the flow cytometric analysis.
Flow cytometry The flow cytometry analyses were essentially done as described earlier (11). In brief, cell suspensions were prepared from inguinal, popliteal, axillary, brachial and cervical lymph nodes from individual mice. Anti-CD4 antibodies used for single stainings were either H129 hybridoma supernatant which was detected by FITC-conjugated goat anti-rat IgG (H + L) (Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA) or phycoerythrin(PE)-conjugated GK.l.5 (Becton Dickinson Immunocytometry System, Mountain View, CA, USA). As a negative control normal rat Ig (NRIG) was employed. Anti-CD4 antibodies remaining on the cell surface were detected by staining the LN cells in vitro with the FITC-conjugated goat antirat IgG(H + L) alone, without the addition of any primary antibody. In double stainings: IM7.8.1 (to CD44) was used as hybridoma supernatant and detected by FITC-conjugated goat anti-rat IgG(H + L) which subsequently was blocked for further binding of rat Ig by an excess
362 . T.
J.
GOLDSCHMIDT, M. ANDERSSON, V. MALMSTROM, and R. HOLMDAHL
of normal rat Ig (NRIG) followed by biotinylated H129 or biotinylated 53.6.7 (to CD8) (19) that was visualized by phycoerythrin-conjugated streptavidin (Becton Dickinson). In the double staining presented in Table 2 were used fluorescein (FITC)-conjugated 53.6.7 (to CD8) and biotinylated 53.7.3 (to CDS) (19), (Becton Dickinson Immunocytometry System). The stained cells were analyzed in a FACStar® (Becton Dickinson). Lymphocytes were selected by forward and right angle scattering and the viability was checked by exclusion of propidium iodide. In single stained samples 104 and in double stained samples cells 2.5-3 x 104 cells were counted.
Proliferation assay The proliferative response of LN cells after immunization with ClI was measured as described earlier with minor modifications (8). Briefly, 14 days after immunization with ClI in the hind footpads and in the base of the tail, the draining LN were removed and cell suspensions prepared. The cells were cultured in triplicates in flat-bottom 96-well plates (Nunc, Roskilde, Denmark) with heat denatured (30 min at 50 DC) rat ClI in the presence and absence of 5 ftg/ml GK.1.5 in DMEM supplemented with Hepes, L-glutamine, penicillin, streptomycin and 1 % fresh normal mouse serum at a concentration of 5 x 106 cells/ml for 96 h in the presence of 1 ftCi (37 kBq) of eH]dThd (Amersham Int., Amersham, Bucks., UK).
Results
CIA development and the levels of anti-CII antibodies in serum DBA/1 male mice received a single anti-CD4 antibody treatment either 10 days before or 5 days after the immunization with rat CII. Two weeks before the experiment the mice were thymectomized to prevent new T cells from maturing after the anti-CD4 treatment was finished. The elimination of CD4+ cells was enhanced by the addition of anti-rat kappa antibodies, as described earlier (11). Anti-CD4 antibodies given before immunization resulted in a delayed onset of arthritis and reduced severity (experiment 1, using DBA/1]), or in a reduced incidence and severity (experiment 2, using DBA/1/0la) (Fig. 1). In the groups treated after the immunization the arthritis development was similar to that in the control group. The levels of anti-ClI antibodies in serum were determined by ELISA. In the control group a strong anti-ClI antibody response was seen on day 14, which further increased, as measured on day 35 (Table 1). In the group treated before immunization, the anti-ClI antibody response was substantially reduced. In the group treated with anti-CD4 after the immunization, no reduction was seen day 14, but the subsequent increase observed day 35 in the control group, did not occur.
Flow cytometric analyses Anti-CD4 treatment 10 days before or 5 days after immunization was found to reduce the number of CD4+ cells to the same extent, as determined on day 14 after immunization (Table 2). The reduction was due to elimination of the CD4 + T cells and not modulation or blockage of CD4 molecules on the cell surface, as indicated by the equal reduction of CD4+ T cells found in the single staining of CD4 and in the double staining of CDS
Activated T Cells not Eliminated by Anti-CD4 Treatment . 363 100
a
control
.,..0--0
c
80 60 40 20 0
d
b !'! 0
6
" III
.2 ~
J:
4
.
2
ic:
III
:;
0 0
10
20
30
40
50
60
a
Days after immunization
10
20
30
40
50
60
Days after immunization
Figure 1. Effects of anti-CD4 treatment on collagen-induced arthritis. DBA/1 male mice, previously thymectomized, were immunized with 50 !!g native rat CII/FCA and received a single anti-CD4 antibody treatment 10 days before or 5 days after immunization. The development of arthritis from two experiments is presented as incidence (a and c) and severity (b and d): Experiment 1, carried out with DBA/1] mice (a and b) and experiment 2, carried out with DBA/1/0la (c and d). The mean arthritic score is not presented for less than 3 arthritic mice. Control mice (0), immunized but otherwise untreated (n = 11 in both exp1 and 2), mice anti-CD4 treatment before immunization (0) (n= 10 in both exp1 and 2) and mice anti-CD4' treatment after immunization (.) (n = 11 in both exp 1 and 2).
and the pan T cell marker CDS, where the CDS+, CDS- population, whether or not the CD4 molecule is modulated, corresponds to the T cells originally expressing CD4. Moreover, staining with the fluorescence antiTable 1. Antibody response to CII after anti-CD4 treatment before or after CII immunization Anti-CD4 treatment'
Untreated Before immunization After immunization
Anti-CII antibodies in serum (!!g/ml) Day 14b
Day 35 C
374 ± 254 81 ± 26 d 297 ± 116
715 ± 375 123 ± 135 e 248 ± 245 1
Thymectomized DBA/1] male mice were given anti-CD4 antibody treatment before (on day -10) or after (on day 5) the immunization (day 0) with native rat CII. b n = 5-6, immunized in both hind foot pads and the base of the tail, in parallel with exp 1. e n = 10-11, immunized in the base of the tail, in exp 1. d P = 0.016, e p < 0.0001, 1 P = 0.012 (Student's t-test).
a
364 . T.
J. GOLDSCHMIDT, M. ANDERSSON, V. MALMSTROM, and R. HOLMDAHL
Table 2. Flow cytometry analyses of lymph node cells from T cell subsets in mice
cn immunized mice cn immunize on day 0
Day 14a
Anti-CD4 treatment
Day S6b
% CD4+
% CDS+, CD8-
% CD4+
Experiment 1 Untreated Before immunizationC After immunizationd
13.6 ± 0.8 4.3 ± 0.5 S.O± 0.6
14.9 ± 1.7 5.2 ± 0.6 5.8 ± 0.9
20.4 ± 4.3 8.4 ± 1.0 7.9 ± 1.6
Experiment 2 Untreated Before immunizationC After immunizationd
15.9/18.0 2.8/5.1 5.8/5.6
18.5120.4 4.6/6.4 7.3/8.0
n.d. e n.d. n.d.
a
b C
d e
Experiment 1: mean ± standard deviation of 3 cell preparations each consisting of LN cells pooled from 2 DBA/IJ mice. Experiment 2: separate values from 2 ceIl preparations each consisting of LN cells pooled from three DBA/1/0Ia mice. Experiment 2: mean of 5 individually stained LN cell suspensions ± standard deviation. 10 days before the immunization on day 0 5 days after the immunization on day 0 n.d. = not done. 60
50
'7 o
40
>
< E c. '-'
b) Before
40 30
20 10
o 60 50
'7
o
>< E c. '-'
c) After
40 30
20 10
o
o
10 rat CII
50 (~g/ml)
250
Figure 2. Proliferative response of LN cells from thymectomized DBA/ IJ mice immunized 14 days earlier with cn and anti-CD4 treated as described in Figure 1. The proliferation to denatured rat cn in vitro of three cell preparations per treatment is presented (each cell preparation derived from two pooled mice): a) control, b) anti-CD4 antibody treated 10 days before the immunization, c) antiCD4 treated 5 days after the immunization. These cell preparations were also subjected to flow cytometric analysis presented in Table 2 as «experiment 1».
Activated T Cells not Eliminated by Anti-CD4 Treatment . 365
rat conjugate only, did not show any blocking H129 antibodies on the cell surface (data not shown). The reduction was long lasting as shown by the analyses day 56. The low relative number of T cells in the untreated but immunized compared to unimmunized mice, is probably due to the relative increase of B cells in LN of immunized individuals.
Proliferative response to ell Aliquotes of the cell suspensions that were analyzed by flow cytometry on day 14 (Table 2) were also tested for in vitro proliferation to denatured rat cn (Fig. 2 and 3). LN cells from mice that were anti-CD4 treated before immunization showed only a minimal response while those treated after immunization exhibited approximately half of (Fig. 2) or the same (Fig. 3a) proliferative capacity when compared to ClI found in the control. Thus the proliferative capacity was retained despite the late treatment. The absence of proliferative response in mice treated before immunization could not be explained by remaining and blocking anti-CD4 antibodies, since no such antibodies could be detected. The proliferation was effectively prevented by the addition of anti-CD4 antibodies to the in vitro culture, indicating that the proliferating cells were CD4+ T cells (Fig. 3). 60
a
50 40
"I
0 ~
>
1 Department of Medical and Physiological Chemistry, University of Uppsala and 2Department of Inflammation Research, Kabi Pharmacia, Uppsala, Sweden
Activated Type II Collagen Reactive T Cells are not Eliminated by in vivo Anti-CD4 Treatment. Implications for Therapeutic Approaches on Autoimmune Arthritis TOM]' GOLDSCHMIDT 1,2, MIKAELANDERSSON 1, VIVIANNEMALMSTROM 1, and RIKARD HOLMDAHL 1 Received October 14, 1991 . Accepted November 18, 1991
Abstract Activation of CD4+ T cells plays an important role in type II collagen (CII) induced arthritis (CIA). The CD4+ T cell dependency is demonstrated by anti-CD4 antibody treatment which suppresses CIA in mice if injected before CII immunization. The same anti-CD4 treatment at a later stage does not suppress CIA, despite extensive elimination of peripheral CD4+ T cells. A possible explanation for this discrepancy is that activated T cells might not be as easily influenced by the anti-CD4 antibodies as resting T cells. To address this question, the proliferative capacity of CII reactive CD4+ lymph node (LN) T cells, in mice treated with antiCD4 antibodies before or after the CII immunization, was analyzed. In mice treated before immunization the capacity of LN cells to proliferate in vitro was markedly suppressed while in mice receiving anti-CD4 treatment after immunization it was retained. Flow cytometric analysis revealed that the anti-CD4 treatment before and after immunization reduced the number of CD4+ LN T cells to the same level. The small population of CD4+ LN cells which were left after anti-CD4 treatment of naive mice all expressed CD44, a marker for previously activated T cells in mice. We propose that activation render CII reactive T cells more resistant to anti-CD4 treatment than virgin T cells are and suggest that the lack of therapeutic effect of late anti-CD4 treatment in CIA does not necessarily implicate that CD4+ T cells are unimportant in that stage of the disease.
Introduction Type II collagen-induced arthritis (CIA) is an experimental tissuespecific autoimmune disease originally characterized in mice and rats which is considered to be an important model for rheumatoid arthritis (RA) (1, 2). CIA is induced by immunization with native type II collagen (CII), which in mice cause arthritis in peripheral joints within three to eight weeks. The pathogenic mechanism involves activation of CD4+ T cells as implicated by the association of arthritis susceptibility to certain MHC class II haplotypes Abbreviations: ClI = type II collagen; CIA = collagen-induced arthritis; RA = rheumatoid arthritis; Ab = antibody; mAb = monoclonal antibody; LN = lymph node; PCA = Freund's complete adjuvant; MHC = major histocompatibility complex
360 . T.
J. GOLDSCHMIDT, M. ANDERSSON, V. MALMSTROM, and R. HOLMDAHL
(3, 4) and the appearance of CD4+ T cells in early stages of arthritis (5). Moreover, the development of CIA can be prevented by injection of antiCD4 antibodies at the time of immunization (6, 7). However, it has been more difficult to affect the disease with anti-CD4 treatment at a later point of time (6,7), when the immune response to cn, shown by the presence of activated T cells and anti-Cn autoantibodies, has already been triggered (8,9). The reason why not anti-CD4 treatment at later stages have been therapeutically effective is unknown. It may be that CD4+ T cells only play a role in the induction of B cells to produce pathogenic autoantibodies. However, mAb treatment to CD4 usually does not result in complete elimination of CD4+ cells (10, 11). An alternative explanation for the absence of therapeutic effects in CIA may be that the monoclonal antibody treatment less efficiently affects activated T cells. We investigated if the occurrence of cn reactive T cells differs in mice which had been anti-CD4 treated before or after immunization with rat CII. The presence of cn reactive T cells was measured by an in vitro proliferation assay and the elimination of CD4 + T cells was analyzed by flow cytometry. The experiments were carried out in CIA susceptible DBA/1 mice in which anti-Cn antibody response and arthritis development were evaluated. Moreover, we studied whether T cells resistant to anti-CD4 or anti-CD8 mAb treatment possessed an activated phenotype by analyzing T cells, left after treatment of unimmunized mice, for the expression of CD44, a cell surface molecule expressed at high density on activated memory T cells (12).
Materials and Methods Induction and evaluation of CIA DBA/I] mice, originally obtained from] ackson Laboratories Inc, were bred and kept at the Biomedical Center in Uppsala and DBA/lOla, derived from Harlan Olac Ltd (Blackthorn, UK), were bred and kept at the Department of Pathology in Uppsala. In all experiments were used 8-12 weeks old, age matched, male mice. Adult thymectomy was done under pentobarbital anesthesia (Mebumal, ACO Likemedel, Goteborg, Sweden) and the mice were allowed to recover two weeks after surgery before the experiments started. Rat cn was prepared by pepsin digestion of a rat chondrosarcoma and subsequent purification with selective salt precipitation and ion exchange chromatography on DEAE-Sepharose (13). For immunization native rat cn was dissolved in 0.1 M acetic acid +4 DC and emulsified 1:1 on ice with Freund's complete adjuvant (FCA) (Difco, Detroit, MI, USA) to a final concentration of 1 mg/m!. Mice that were used in the arthritis experiment were injected intradermally at the base of the tail with 50 [11 containing 50 [1g of rat cn emulsified 1:1 in FCA. Mice that were used for the proliferation and flow cytometric analyses were immunized intradermally at the base of the tail and both hind foot pads within all 150 [1g rat cn. Arthritis development in the four paws was judged by a macroscopic scoring system ranging from 0 to 3 (1 = swelling and/or redness of one toe or finger joint, 2 = two or more joints involved and 3 = severe arthritis in the entire paw).
Activated T Cells not Eliminated by Anti-CD4 Treatment . 361
Quantification of anti-CIl antibodies in serum Sera were collected individually and stored at -20 DC until assayed. For the quantification of anti-Cn reactive autoantibodies in serum, a modified standard ELISA technique was used (4). In brief, micro-ELISA plates (Dynatech, Plochingen, Germany) were coated with 10 [Lg/mlof native rat cn. All tests were carried out in duplicates, and the standard deviations did not exceed 10 %. The amount of bound antibody was estimated after incubation with a goat antimouse IgG (H + L) affinity purified antibody conjugated to alkaline phosphatase 0ackson Immunoresearch Laboratories, West Grove, PA, USA). The subsequent quantification of bound enzyme was performed with a paranitrophenol containing substrate buffer in a Titertek Multiscan® spectrophotometer. To estimate the amount of anti-Cn reactive antibodies present in the serum samples, an affinity purified DBA/1 anti-Cn antibody preparation was used as standard and titrated in parallel with the unknown serum samples.
Monoclonal antibodies The following hybridomas were cultured; H129.19 (H129), which produce a rat IgG2a antibody specific for mouse CD4 (14); GKl.5, which produce a rat IgG2b which binds to mouse CD4 in a competitive fashion to H129 (15); MAR18:5, which produce a mouse IgG2a specific for the rat kappa chain (16); IM7.8.1, which produce a rat IgG2b to mouse CD44, was kindly provided by Dr. ROBSON MCDONALD, Lausanne (17). H129 hybridoma supernatant was concentrated and partly purified by ammonium sulphate precipitation between 25 and 45 percentage saturation, dialyzed to PBS, sterile filtered and stored at -20 DC until use. The specific H129 content was estimated by ELISA with an affinity purified H129 preparation as standard (4). GK1.5 was chromatographically purified (18). MARl.8 was affinity purified on Protein A-Sepharose (Pharmacia AB, Uppsala, Sweden), dialyzed with PBS and sterile filtered. The protein content was quantified by absorbance measurement at 280 nm.
Anti-CD4 antibody treatment in vivo The anti-CD4 treatment was performed by injecting 100 [Lg H129 i.p., which one hour later was followed by 200 [Lg MAR18.5 i.p. In this sandwich method the anti-rat kappa antibody MAR18.5 binds to H129 in vivo and enhances the elimination of CD4+ cells as earlier described (11). In the arthritis and proliferation experiments the anti-CD4 treatment was given either 10 days before or 5 days after immunization of the previously thymectomized mice. Before the immunization, the pre-treated mice were tested for the presence of persisting antiCD4 (H129) rat antibodies in serum. Micro-ELISA plates (Dynatech) were coated with goat anti-rat Ig (H + L) which had been mouse serum absorbed 0ackson Immunoresearch Laboratories, Inc.) and the serum titrated as described above. Bound rat monoclonal antibodies were subsequently detected by alkaline phosphatase conjugated, mouse serum absorbed, goat anti-rat Ig (H + L) 0ackson Immunoresearch Laboratories, Inc.). Purified H129 was used as standard and the calculations were performed as described above for anticn antibodies. In the experiment where the expression of CD44 on LN cells was measured, naive DBA/1] mice were anti-CD4 treated 8 days prior to the flow cytometric analysis.
Flow cytometry The flow cytometry analyses were essentially done as described earlier (11). In brief, cell suspensions were prepared from inguinal, popliteal, axillary, brachial and cervical lymph nodes from individual mice. Anti-CD4 antibodies used for single stainings were either H129 hybridoma supernatant which was detected by FITC-conjugated goat anti-rat IgG (H + L) (Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA) or phycoerythrin(PE)-conjugated GK.l.5 (Becton Dickinson Immunocytometry System, Mountain View, CA, USA). As a negative control normal rat Ig (NRIG) was employed. Anti-CD4 antibodies remaining on the cell surface were detected by staining the LN cells in vitro with the FITC-conjugated goat antirat IgG(H + L) alone, without the addition of any primary antibody. In double stainings: IM7.8.1 (to CD44) was used as hybridoma supernatant and detected by FITC-conjugated goat anti-rat IgG(H + L) which subsequently was blocked for further binding of rat Ig by an excess
362 . T.
J.
GOLDSCHMIDT, M. ANDERSSON, V. MALMSTROM, and R. HOLMDAHL
of normal rat Ig (NRIG) followed by biotinylated H129 or biotinylated 53.6.7 (to CD8) (19) that was visualized by phycoerythrin-conjugated streptavidin (Becton Dickinson). In the double staining presented in Table 2 were used fluorescein (FITC)-conjugated 53.6.7 (to CD8) and biotinylated 53.7.3 (to CDS) (19), (Becton Dickinson Immunocytometry System). The stained cells were analyzed in a FACStar® (Becton Dickinson). Lymphocytes were selected by forward and right angle scattering and the viability was checked by exclusion of propidium iodide. In single stained samples 104 and in double stained samples cells 2.5-3 x 104 cells were counted.
Proliferation assay The proliferative response of LN cells after immunization with ClI was measured as described earlier with minor modifications (8). Briefly, 14 days after immunization with ClI in the hind footpads and in the base of the tail, the draining LN were removed and cell suspensions prepared. The cells were cultured in triplicates in flat-bottom 96-well plates (Nunc, Roskilde, Denmark) with heat denatured (30 min at 50 DC) rat ClI in the presence and absence of 5 ftg/ml GK.1.5 in DMEM supplemented with Hepes, L-glutamine, penicillin, streptomycin and 1 % fresh normal mouse serum at a concentration of 5 x 106 cells/ml for 96 h in the presence of 1 ftCi (37 kBq) of eH]dThd (Amersham Int., Amersham, Bucks., UK).
Results
CIA development and the levels of anti-CII antibodies in serum DBA/1 male mice received a single anti-CD4 antibody treatment either 10 days before or 5 days after the immunization with rat CII. Two weeks before the experiment the mice were thymectomized to prevent new T cells from maturing after the anti-CD4 treatment was finished. The elimination of CD4+ cells was enhanced by the addition of anti-rat kappa antibodies, as described earlier (11). Anti-CD4 antibodies given before immunization resulted in a delayed onset of arthritis and reduced severity (experiment 1, using DBA/1]), or in a reduced incidence and severity (experiment 2, using DBA/1/0la) (Fig. 1). In the groups treated after the immunization the arthritis development was similar to that in the control group. The levels of anti-ClI antibodies in serum were determined by ELISA. In the control group a strong anti-ClI antibody response was seen on day 14, which further increased, as measured on day 35 (Table 1). In the group treated before immunization, the anti-ClI antibody response was substantially reduced. In the group treated with anti-CD4 after the immunization, no reduction was seen day 14, but the subsequent increase observed day 35 in the control group, did not occur.
Flow cytometric analyses Anti-CD4 treatment 10 days before or 5 days after immunization was found to reduce the number of CD4+ cells to the same extent, as determined on day 14 after immunization (Table 2). The reduction was due to elimination of the CD4 + T cells and not modulation or blockage of CD4 molecules on the cell surface, as indicated by the equal reduction of CD4+ T cells found in the single staining of CD4 and in the double staining of CDS
Activated T Cells not Eliminated by Anti-CD4 Treatment . 363 100
a
control
.,..0--0
c
80 60 40 20 0
d
b !'! 0
6
" III
.2 ~
J:
4
.
2
ic:
III
:;
0 0
10
20
30
40
50
60
a
Days after immunization
10
20
30
40
50
60
Days after immunization
Figure 1. Effects of anti-CD4 treatment on collagen-induced arthritis. DBA/1 male mice, previously thymectomized, were immunized with 50 !!g native rat CII/FCA and received a single anti-CD4 antibody treatment 10 days before or 5 days after immunization. The development of arthritis from two experiments is presented as incidence (a and c) and severity (b and d): Experiment 1, carried out with DBA/1] mice (a and b) and experiment 2, carried out with DBA/1/0la (c and d). The mean arthritic score is not presented for less than 3 arthritic mice. Control mice (0), immunized but otherwise untreated (n = 11 in both exp1 and 2), mice anti-CD4 treatment before immunization (0) (n= 10 in both exp1 and 2) and mice anti-CD4' treatment after immunization (.) (n = 11 in both exp 1 and 2).
and the pan T cell marker CDS, where the CDS+, CDS- population, whether or not the CD4 molecule is modulated, corresponds to the T cells originally expressing CD4. Moreover, staining with the fluorescence antiTable 1. Antibody response to CII after anti-CD4 treatment before or after CII immunization Anti-CD4 treatment'
Untreated Before immunization After immunization
Anti-CII antibodies in serum (!!g/ml) Day 14b
Day 35 C
374 ± 254 81 ± 26 d 297 ± 116
715 ± 375 123 ± 135 e 248 ± 245 1
Thymectomized DBA/1] male mice were given anti-CD4 antibody treatment before (on day -10) or after (on day 5) the immunization (day 0) with native rat CII. b n = 5-6, immunized in both hind foot pads and the base of the tail, in parallel with exp 1. e n = 10-11, immunized in the base of the tail, in exp 1. d P = 0.016, e p < 0.0001, 1 P = 0.012 (Student's t-test).
a
364 . T.
J. GOLDSCHMIDT, M. ANDERSSON, V. MALMSTROM, and R. HOLMDAHL
Table 2. Flow cytometry analyses of lymph node cells from T cell subsets in mice
cn immunized mice cn immunize on day 0
Day 14a
Anti-CD4 treatment
Day S6b
% CD4+
% CDS+, CD8-
% CD4+
Experiment 1 Untreated Before immunizationC After immunizationd
13.6 ± 0.8 4.3 ± 0.5 S.O± 0.6
14.9 ± 1.7 5.2 ± 0.6 5.8 ± 0.9
20.4 ± 4.3 8.4 ± 1.0 7.9 ± 1.6
Experiment 2 Untreated Before immunizationC After immunizationd
15.9/18.0 2.8/5.1 5.8/5.6
18.5120.4 4.6/6.4 7.3/8.0
n.d. e n.d. n.d.
a
b C
d e
Experiment 1: mean ± standard deviation of 3 cell preparations each consisting of LN cells pooled from 2 DBA/IJ mice. Experiment 2: separate values from 2 ceIl preparations each consisting of LN cells pooled from three DBA/1/0Ia mice. Experiment 2: mean of 5 individually stained LN cell suspensions ± standard deviation. 10 days before the immunization on day 0 5 days after the immunization on day 0 n.d. = not done. 60
50
'7 o
40
>
< E c. '-'
b) Before
40 30
20 10
o 60 50
'7
o
>< E c. '-'
c) After
40 30
20 10
o
o
10 rat CII
50 (~g/ml)
250
Figure 2. Proliferative response of LN cells from thymectomized DBA/ IJ mice immunized 14 days earlier with cn and anti-CD4 treated as described in Figure 1. The proliferation to denatured rat cn in vitro of three cell preparations per treatment is presented (each cell preparation derived from two pooled mice): a) control, b) anti-CD4 antibody treated 10 days before the immunization, c) antiCD4 treated 5 days after the immunization. These cell preparations were also subjected to flow cytometric analysis presented in Table 2 as «experiment 1».
Activated T Cells not Eliminated by Anti-CD4 Treatment . 365
rat conjugate only, did not show any blocking H129 antibodies on the cell surface (data not shown). The reduction was long lasting as shown by the analyses day 56. The low relative number of T cells in the untreated but immunized compared to unimmunized mice, is probably due to the relative increase of B cells in LN of immunized individuals.
Proliferative response to ell Aliquotes of the cell suspensions that were analyzed by flow cytometry on day 14 (Table 2) were also tested for in vitro proliferation to denatured rat cn (Fig. 2 and 3). LN cells from mice that were anti-CD4 treated before immunization showed only a minimal response while those treated after immunization exhibited approximately half of (Fig. 2) or the same (Fig. 3a) proliferative capacity when compared to ClI found in the control. Thus the proliferative capacity was retained despite the late treatment. The absence of proliferative response in mice treated before immunization could not be explained by remaining and blocking anti-CD4 antibodies, since no such antibodies could be detected. The proliferation was effectively prevented by the addition of anti-CD4 antibodies to the in vitro culture, indicating that the proliferating cells were CD4+ T cells (Fig. 3). 60
a
50 40
"I
0 ~
>
