Collagen-Induced Arthritis CD28-Deficient Mice Are Highly

CD28-Deficient Mice Are Highly Resistant to Collagen-Induced Arthritis Yoshifumi Tada, Kohei Nagasawa, Alexandra Ho, Fumiaki Morito, Osamu Ushiyama, Noriaki Suzuki, Hideaki Ohta and This information is current as Tak Wah Mak of September 24, 2021. J Immunol 1999; 162:203-208; ; http://www.jimmunol.org/content/162/1/203 Downloaded from

References This article cites 34 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/162/1/203.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 24, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. CD28-Deﬁcient Mice Are Highly Resistant to Collagen-Induced Arthritis1

Yoshifumi Tada,* Kohei Nagasawa,* Alexandra Ho,† Fumiaki Morito,* Osamu Ushiyama,* Noriaki Suzuki,* Hideaki Ohta,* and Tak Wah Mak†

CD28 provides a critical costimulatory signal in Ag-specific T cell activation. Recent studies have revealed an important role for CD28 in the development of autoimmune diseases. We have examined the role of CD28 in collagen-induced arthritis (CIA) by inducing CIA in CD28-deficient DBA/1 mice. CD28-deficient mice never developed arthritis and showed markedly decreased levels ,of IgG and IgM anti-type II collagen (CII) Abs. In addition, the CD28؉/؊ mice had similar levels of IgG1 and IgG2a anti-CII Abs whereas in the CD28-deficient mice the level of IgG1 anti-CII Abs was decreased compared with that of IgG2a. IFN-␥ production by lymph node cells in response to CII was also reduced. CD28-deficient mice were either immunized four times with CII in CFA to augment Ag loading or given low doses of IL-12 to enhance Th1 type responses. Both treatments resulted in a very low incidence Downloaded from of CIA development and minimal disease. CD28-deficient mice developed arthritis from injection of lymph node cells from CII-immunized wild-type mice, followed by immunization with CII in CFA. Taken together, these results indicate that costimulation of CD28 cannot be replaced by repeated activation through TCR or other costimulatory molecules. Thus, CD28 plays a critical role in both cellular and humoral immunity against CII and is indispensable for the development of CIA. The Journal of Immunology, 1999, 162: 203–208. http://www.jimmunol.org/ xtensive investigations have recently attempted to clarify impaired, IgG Ab production in response to peptides or virus is the role of costimulatory signals in the activation of naive reduced, and serum IgG1 levels are decreased (17). However, anal- E T cells. Costimulatory signals are provided by the phys- yses of various autoimmune disease models using CD28-deficient ical interaction between surface molecules expressed on T cells mice have shown different and contradictory results. NOD mice and those on APCs (1, 2). CD28 has been shown to be one of the lacking CD28 unexpectedly exhibited accelerated development of most important costimulatory molecules for T cell activation (3– diabetes (18). Autoimmune myocarditis was induced with similar 5), making the manipulation of CD28-mediated activation path- incidence in CD28-deficient and wild-type mice, although the se- ways an important issue in both the regulation of T cell activation verity was decreased in the mutants (19). These studies indicate by guest on September 24, 2021 and the treatment of pathologic conditions. CTLA4-Ig has been that autoantigen-specific and pathologic T cells can be induced, at widely used to block the binding of CD28 on T cells to its ligands least in certain cases, in the absence of CD28. It was proposed that CD80 or CD86 on APCs, resulting in the inhibition of costimula- the absence of CD28 signaling might be overcome by other co- tory signals (6). In in vivo studies, it has been shown that stimulatory molecules or that CD28 costimulatory signals might CTLA4-Ig treatment enhances the survival of pancreatic islet not be absolutely required in some circumstances, such as where xenografts (7) and cardiac allografts (8, 9) and reduces the severity there are high levels of cytokines, or in particular T cell subsets. of graft-vs-host disease (10). Furthermore, in autoimmune disease In this report, we have investigated the role of CD28 in the CIA models such as murine lupus (11), nonobese diabetes (NOD)3 (12), model using CD28-deficient mice. The results show that CD28- experimental allergic encephalomyelitis (13, 14), and collagen-in- deficient mice are highly resistant to the induction of CIA, even duced arthritis (CIA) (15, 16), disease has been prevented or suc- when reinforced immunization protocols are applied. The minimal cessfully treated by CTLA4-Ig administration. These results imply response of T cells to type II collagen (CII) and the low IgM and a critical role for CD28 in the development of these diseases and IgG anti-CII Ab levels indicate that both cellular and humoral suggest that the inhibition of CD28-mediated activation pathways immunity against CII in CIA are highly dependent on costimula- can be a useful strategy in their amelioration. tory signaling through CD28. Studies of CD28Ϫ/Ϫ gene-targeted mice have revealed that in the absence of CD28, T cell responses to mitogens in vitro are Materials and Methods Mice † *Department of Internal Medicine, Saga Medical School, Saga, 849, Japan; and Am- CD28-deficient mice (19) were backcrossed into the DBA/1 background gen Institute, Ontario Cancer Institute and Departments of Medical Biophysics and (H-2q) for four generations and typed by PCR using ear biopsy-derived Immunology, University of Toronto, Toronto, Ontario, Canada Ϫ Ϫ ϩ Ϫ ϩ ϩ DNA (19). In all experiments, only CD28 / , CD28 / , and CD28 / Received for publication February 19, 1998. Accepted for publication September littermates were used. All mice were 10 to 16 wk of age at the time of 16, 1998. immunization. Animals were maintained in the Ontario Cancer Institute The costs of publication of this article were defrayed in part by the payment of page animal facility or the Saga Medical School animal facility. Care of animals charges. This article must therefore be hereby marked advertisement in accordance was in accordance with the guidelines of the Canadian Medical Research with 18 U.S.C. Section 1734 solely to indicate this fact. Council and the Saga Medical School guidelines for animal experimentation. 1 This work was supported by the Medical Research Council of Canada. Induction of CIA 2 Address correspondence and reprint requests to Dr. Tak Wah Mak, Amgen Institute, 620 University Ave., Toronto, Ontario, Canada M5G 2 M9. Mice were immunized intradermally at the base of the tail with 150 ␮gof 3 Abbreviations used in this paper: NOD, nonobese diabetes; CIA, collagen-induced bovine CII (Elastin Products, Owensville, MO) emulsified with an equal arthritis; CII, collagen type II; dCII, denatured collagen type II; LN, lymph node. volume of CFA, containing 200 ␮g of H37RA Mycobacterium tuberculosis

(Difco, Detroit, MI). Mice were boosted by intradermal injection with 150 Table I. Collagen-induced arthritis in CD28ϩ/ϩ, CD28ϩ/Ϫ, and ␮g of bovine CII in CFA (IFA, Difco) on day 21. In one experiment, mice CD28Ϫ/Ϫ DBA/1 mice were immunized four times with CII in CFA to augment Ag priming. To address the role of IL-12 on the development of arthritis in CD28-deficient Genotype Incidence of Disease Arthritis Indexa mice, mice were injected i.p. with 10 or 100 ng IL-12 on days 1–5 and 21–25 after the initial challenge (20). Arthritis development was checked Expt. 1 by inspection three times a week, and the inflammation of four paws was CD28ϩ/ϩ 6/8 (75.0%) 4.3 Ϯ 0.9 graded from 0 to 3 as described previously (21, 22). Each paw was graded CD28ϩ/Ϫ 6/8 (75.0%) 4.3 Ϯ 1.1 and the four scores were added such that the maximal score per mouse was CD28Ϫ/Ϫ 0/5 (0.0%) 12. The arthritis index was calculated by dividing the total score of the Expt. 2 experimental mice by the number of arthritic mice or the total number of CD28ϩ/Ϫ 8/10 (80.0%) 6.6 Ϯ 0.4 mice. CD28Ϫ/Ϫ 0/8 (0.0%) Expt. 3 Measurement of serum anti-CII Ab levels CD28ϩ/Ϫ 6/8 (75.0%) 5.8 Ϯ 1.9 Ϫ/Ϫ The level of serum Abs to CII was measured by ELISA as described (22). CD28 0/6 (0.0%) Briefly, microtiter plates (Maxisorp, Nunc, Denmark) were coated with a Mean Ϯ SEM of the maximal scores of arthritic mice. native bovine CII at 10 ␮g/ml overnight at 4°C. After washing, serum samples were added in serial dilution and incubated for1hat37°C. After four washes, peroxidase-conjugated goat anti-mouse IgG (Biosource In- ternational, Camarillo, CA), IgG1, or IgG2a (Southern Biotechnology As- ient mice were then immunized with CII and CFA on the next day, and sociates, Birmingham, AL) was added and incubated for1hat37°C. Ab arthritis development was observed. binding was visualized using o-phenylenediamine (Sigma, St. Louis, MO). Downloaded from A standard serum composed of a mixture of sera from arthritic mice was Results added to each plate in serial dilutions and a standard curve was constructed. Clinical course of CIA and anti-CII Ab levels in CD28Ϫ/Ϫ DBA/ The standard serum was defined as 100 U and Ab titers of serum samples 1 mice were calculated from the standard curve. CD28Ϫ/Ϫ, CD28ϩ/Ϫ, and CD28ϩ/ϩ littermate mice were immu- Histologic examination nized with CII and observed for signs of arthritis. As shown in ϩ/Ϫ Ϫ/Ϫ Ϫ/Ϫ Joint tissues from CD28 and CD28 DBA/1 mice were excised 9 to Table I, CD28 DBA/1 mice never developed arthritis in 3 ex- http://www.jimmunol.org/ 10 wk after immunization and fixed in 10% buffered formalin, decalcified periments using a total of 19 mutant mice. CD28ϩ/Ϫ mice and in 10% EDTA, embedded in paraffin, sectioned, and stained with hema- CD28ϩ/ϩ mice developed arthritis with similar incidence and se- toxylin and eosin. verity (Table I and Fig. 1). Histologic examination revealed that in Cytokine production of lymph node cells contrast to the typical arthritis characterized by the dense cellular ϩ/ϩ ␥ infiltration and bone erosion observed in the joints of CD28 IFN- , IL-4, and IL-10 production by LN cells was measured as previously ϩ/Ϫ Ϫ/Ϫ described (22). Fourteen days after immunization, LN (inguinal, paraaortic, and CD28 mice, all joints examined in five CD28 mice axillary, and popliteal) were removed, and the cells were resuspended in showed no signs of inflammation. Ϫ DMEM supplemented with 5 ϫ 10 5 M 2-ME, 20 mM HEPES, and 4% IgG and IgM anti-CII Ab levels were determined on days 28, 42, 6 ϫ by guest on September 24, 2021 autologous mouse serum. Cells were seeded at 4 10 /well in 96-well and 56. As shown in Fig. 2, A and C, anti-CII Ab levels of IgG and flat-bottom microtiter plates (Nunc) and stimulated with denatured bovine Ϫ/Ϫ CII (dCII) at 200 ␮g/ml for 48 or 72 h. Cytokines produced in the culture IgM isotypes were markedly decreased in CD28 mice as com- ϩ/ϩ ϩ/Ϫ supernatant was measured using an ELISA kit (IFN-␥; Genzyme, Cam- pared with CD28 and CD28 mice throughout the course of bridge, MA; IL-4 and IL-10: Biosource International) according to the CIA. IgG anti-CII Ab levels were similar between CD28ϩ/ϩ and manufacturer’s instructions. CD28ϩ/Ϫ mice (Fig. 2A). It has been shown that CD28Ϫ/Ϫ mice RT-PCR have reduced serum IgG1 levels (17) and reduced production of IgG1 Ab in response to Ag priming (19). We therefore evaluated ϩ/Ϫ Ϫ/Ϫ CD28 DBA/1 and CD28 DBA/1 mice were immunized with CII the levels of IgG1 and IgG2a subclass Abs against CII in CD28ϩ/Ϫ and CFA and draining LN were removed after 14 or 21 days. Total RNA Ϫ/Ϫ Ϫ/Ϫ was extracted with TRI-ZOL reagent (Life Technologies, Gaithersburg, and CD28 mice (Fig. 2B). CD28 mice showed markedly Ϫ/Ϫ MD). cDNA was obtained by reverse transcription of 4 ␮g of total RNA decreased levels of both subclasses. However, in CD28 mice, using oligo(dT) and a RT-PCR kit (Pharmacia Biotech, Uppsala, Sweden). IgG2a anti-CII levels were significantly higher than IgG1 anti-CII PCR was done by four different conditions to compare detectable level of levels, whereas both subclasses showed similar levels in CD28ϩ/ϩ each samples, namely using 12 or 36% of the reaction sample as the tem- and CD28ϩ/Ϫ mice. These results imply that the CD28 costimu- plate and with 30 or 35 cycles. Primers used were as follows: ␤-actin, GTGGGCCGCTCTAGGCACCAAandCTCTTTGATGTCACGCACGAT latory molecule is critical in IgG and IgM anti-CII Ab production TTC; IFN-␥, TGCATCTTGGCTTTGCAGCTCTTCCTCATGGC and TGGACCTGTGGGTTGTTGACCTCAAACTTGGC; IL-4, ATGGGTCT CAACCCCCAGCTAGT and GCTCTTTAGGCTTTCCAGGAAGTC. Reaction conditions were the following: 94°C for 40 s; 55°C for 60 s; 72°C for 90 s. PCR products were separated on 2% agarose gels and visualized with ethidium bromide. Adoptive transfer of arthritis DBA/1 mice were immunized with CII and CFA. After 14 days, LN were removed and cell suspensions were prepared in HBSS. In some experiments, B cells were further depleted by magnetic cell sorting with anti- mouse CD45R Ab (RA3-6B2, Cedarlane, Hornby, Canada) coupled to magnetic beads (Biomag, Perseptive Biosystems, Framingham, MA). LN ϫ 6 cells were resuspended in DMEM-5% FCS at 2 10 cells/ml and mixed FIGURE 1. Development of CIA in CD28-deficient mice. CD28ϩ/ϩ with beads previously coupled with anti-CD45R Abs at a cell-beads ratio ϩ/Ϫ Ϫ/Ϫ of 1:50. The cell suspensions were incubated for 30 min on ice, and B cells DBA/1, CD28 DBA/1, and CD28 DBA/1 mice were immunized bound to beads were magnetically separated. The resultant cell population with bovine CII in CFA, and signs of arthritis were monitored as described contained Ͼ93% TCR-␣␤ϩ cells and Ͻ3% surface IgMϩ cells as deter- in Materials and Methods. The arthritis index was calculated from the first Ϫ Ϫ mined by flow cytometric analysis. Whole LN cells or purified T cells were experiment described in Table I. No CD28 / DBA/1 mice developed resuspended in HBSS and injected i.p. into CD28Ϫ/Ϫ DBA/1 mice. Recip- arthritis. The Journal of Immunology 205

FIGURE 2. IgG and IgM anti-CII Ab levels in CD28ϩ/ϩ DBA/1, CD28ϩ/Ϫ DBA/1, and CD28Ϫ/Ϫ DBA/1 mice. Anti-CII IgG Ab (A) and anti-CII IgM Ab (B) levels were measured on days 28, 42, and 56 by ELISA. Anti-CII levels of CD28Ϫ/Ϫ mice were signiﬁcantly decreased as compared with those of CD28ϩ/ϩ and CD28ϩ/Ϫ mice (p Ͻ 0.01). IgG1 and IgG2a isotypes of anti-CII Abs were measured on day 42 sera by ELISA (C). IgG1 levels were signiﬁcantly Downloaded from lower than IgG2a levels (p Ͻ 0.01). Mean Ϯ SEM are shown. http://www.jimmunol.org/

and in the development of CIA and that the production of IgG1 CD28Ϫ/Ϫ DBA/1 mice developed CIA following repeated immu- anti-CII Ab was severely depressed in CD28Ϫ/Ϫ mice. nizations. Mice were immunized with CII in CFA four times (on days 0, 16, 32, and 48). However, only 1 of 13 CD28Ϫ/Ϫ mice Cytokine production and mRNA expression by lymph node cells developed mild and transient arthritis (grade 1, from day 56 to day by guest on September 24, 2021 ϩ/ϩ ϩ/Ϫ Ϫ/Ϫ from CD28 , CD28 and CD28 DBA/1 mice 65, Table II). IgG anti-CII Ab levels of CD28Ϫ/Ϫ mice were in- To analyze Ag-specific T cell responses in CD28Ϫ/Ϫ DBA/1 mice, creased compared with those of mice immunized with a regular IFN-␥ production by LN cells in response to dCII was examined. Mice were killed on day 14 after immunization and numbers of viable LN cells were counted. Cells were stimulated with dCII, and IFN-␥ in the supernatant was measured. Results from two repre- sentative experiments are shown in Fig. 3. LN cells from CD28Ϫ/Ϫ mice produced a markedly decreased amount of IFN-␥ in response to dCII compared with that produced by cells from CD28ϩ/Ϫ or CD28ϩ/ϩ mice. Furthermore, the total number of LN cells recov- ered from CD28Ϫ/Ϫ mice was ϳ50% that of control mice (Fig. 3). We also measured IL-4 and IL-10 production by LN cells. How- ever, these Th2 cytokines were undetectable in culture supernatant of LN cells stimulated with dCII from CD28Ϫ/Ϫ as well as CD28ϩ/Ϫ mice. To further investigate the Th1 and Th2 cytokine expression, mRNA levels of IFN-␥ and IL-4 were examined in the LN of CD28Ϫ/Ϫ and CD28ϩ/Ϫ DBA/1 mice after immunization with CII. As shown in Fig. 4, no differences were observed in the expression of IFN-␥ in CD28Ϫ/Ϫ and CD28ϩ/Ϫ mice, and although one CD28Ϫ/Ϫ mouse showed decreased level, IL-4 expression could be detected in CD28Ϫ/Ϫ as well as CD28ϩ/Ϫ mice. These results suggest that the expansion and IFN-␥ production of LN cells in response to CII is defective in CD28Ϫ/Ϫ mice. FIGURE 3. Total cell number and IFN-␥ production of LN cells from CD28ϩ/ϩ DBA/1, CD28ϩ/Ϫ DBA/1, and CD28Ϫ/Ϫ DBA/1 mice. Mice Reinforced induction of CIA in CD28Ϫ/Ϫ DBA/1 mice were immunized with CII in CFA. On day 14, the draining LNs were removed and the total cell numbers were counted (hatched column). Ad- Previous studies have shown that in some cases, the absence of justed numbers of LN cells were stimulated with denatured CII (0.2 mg/ml) costimulation via CD28 can be overcome by prolonged stimulation (open column) or medium only (filled column) for 2 days, and IFN-␥ in the Ϫ/Ϫ through the TCR (23) and that CD28 T cells can initiate but not supernatant was measured by ELISA. The results presented are the average sustain a primary Ag-specific response (24). We examined whether of two mice per group. Two independent experiments are shown. 206 COLLAGEN-INDUCED ARTHRITIS IN CD28-DEFICIENT MICE

FIGURE 4. Expression of IFN-␥ and IL-4 mRNA in LNs from CII- immunized CD28ϩ/Ϫ DBA/1 and CD28Ϫ/Ϫ DBA/1 mice. Draining LNs were removed 14 days after the immunization. Total RNA was extracted and RT-PCR was conducted using primers for IFN-␥, IL-4, and ␤-actin. ϩ Ϫ Ϫ Ϫ Lanes 1-3, CD28ϩ/Ϫ DBA/1 mice; lanes 4–7, CD28Ϫ/Ϫ DBA/1 mice. PCR FIGURE 5. Anti-CII IgG Ab levels in CD28 / DBA/1 and CD28 / products were IFN-␥ 365 bp, IL-4 399 bp, ␤-actin 540 bp. PCR conditions DBA/1 mice which were repeatedly immunized. Mice were immunized (percent volume of cDNA sample used and cycles) were as follows: IFN-␥, with CII in CFA four times, and serum levels of anti-CII IgG Ab levels 36% and 35 cycles; IL-4, 12% and 35 cycles; ␤-actin, 12% and 30 cycles. were measured on days 26 (after second immunization), 42 (after third immunization), and 59 (after fourth immunization). Open circles show the Ab levels of the single arthritic CD28Ϫ/Ϫ DBA/1 mouse. Ϫ/Ϫ

protocol (Figs. 2 and 5). The one CD28 mouse that developed Downloaded from Ϫ/Ϫ arthritis showed the highest anti-CII level among all CD28 ϩ/ϩ mice (Fig. 5). IgG subclass anti-CII Ab levels showed that IgG1 of T cells and B cells from CII-primed CD28 mice and that T anti-CII Abs were lower than IgG2a Abs in CD28Ϫ/Ϫ mice (day cells only may transfer arthritis but with less efficacy. 42, IgG1 2.0 Ϯ 0.8 U, IgG2a 20.6 Ϯ 7.0 U, p Ͻ 0.001), whereas both Ab levels were not different in CD28ϩ/Ϫ mice (IgG1 92.4 Ϯ Discussion 16.4 U, IgG2a 101.0 Ϯ 29.3 U). These results indicate that In this report, we show that CD28-deficient DBA/1 mice do not Ϫ Ϫ http://www.jimmunol.org/ CD28 / mice are still resistant to the development of CIA even develop arthritis after standard immunization with CII. CD28-de- after repeated immunization with CII plus CFA. ficient mice showed markedly depressed IgG anti-CII Ab levels We also examined the effect of the administration of low doses and low production of IFN-␥ by LN cells in response to CII. We Ϫ Ϫ of IL-12 on the development of CIA in CD28 / mice. Germann also observed low IgM anti-CII Ab levels in CD28-deficient mice, et al. (20) showed that the low incidence of CIA in DBA/1 mice in contrast to previous reports where serum IgM levels and IgM immunized with CII in IFA (instead of CFA) was greatly increased Ab production against vesicular stomatitis virus were normal (19). when IL-12 was coadministered, suggesting the importance of Th1 Furthermore, CD28ϩ/Ϫ mice showed mild decreased IgM anti-CII type responses in priming with CII. However, in our hands, Ab levels as compared with those of CD28ϩ/ϩ mice (Fig. 2B). Ϫ/Ϫ CD28 DBA/1 mice showed only a minimal response after the These data suggest that IgM Ab production, at least to some Ags by guest on September 24, 2021 addition of low doses of IL-12; only one mouse of six developed or in certain genetic background, may be regulated by CD28. Our mild and transient disease at 10 ng of IL-12 (grade 1, from day 39 results indicate that both the cellular and humoral immune re- to day 67, Table II). IgG anti-CII Ab levels were not increased in sponses to CII are regulated by CD28 costimulation. They agree Ϫ Ϫ IL-12-injected CD28 / mice as compared with levels in mice with previous studies of CIA in rats (15) and mice (16), in both of that underwent the standard immunization protocol (data not which the administration of CTLA4-Ig blocked development of shown). These results indicate that IL-12 fails to enhance anti-CII arthritis and significantly reduced anti-CII Ab levels and T cell Ϫ Ϫ Ab production or induce arthritis in CD28 / DBA/1 mice. response to CII. They also agree with studies of various autoimmune disease models using CTLA4-Ig (11–14, 25, 26) but differ Adoptive transfer of arthritis by DBA/1 lymph node cells to Ϫ/Ϫ from some studies in which CD28-deficient mice were used (18, CD28 DBA/1 mice 19). There may be a fundamental difference between the congen- To directly investigate whether the resistance of CD28Ϫ/Ϫ mice to ital loss of CD28 and treatment with CTLA4-Ig, or the discrepancy CIA depends on the function of lymphoid cells, LN cells, or pu- may be due to variation among disease models. Functional redun- rified T cells from CII-immunized DBA/1 mice were injected into dancy often exists in knockout mice, and the role of a particular naive CD28Ϫ/Ϫ mice, then recipient mice were immunized with molecule can be compensated for or overcome by other molecules. CII. CD28Ϫ/Ϫ mice transferred with whole LN cells from wild- For example, we have previously shown that CD4-deficient mice type mice developed moderate arthritis after 6 to 9 days and lasted develop CIA with normal incidence and severity and that TCR- for Ͼ10 days, on the other hand, mice transferred purified T cells ␣␤ϩCD4ϪCD8Ϫ T cells can replace the role of CD4ϩ T cells in show low incidence and mild arthritis (Table III). These results these mice (21). This result contrasts with the demonstration that indicate that CD28Ϫ/Ϫ mice can develop arthritis in the presence anti-CD4 treatment abolishes development of CIA in normal mice

Table II. Collagen-induced arthritis with modiﬁed induction in CD28Ϫ/Ϫ DBA/1 mice

Repeated Immunization IL-12 Administration

IL-12 dose Genotype Incidence of arthritis Arthritis indexa (ng) Incidence of arthritis Arthritis index

CD28ϩ/Ϫ 8/9 (88.9%) 5.5 Ϯ 0.7 8/11 (72.7%) 6.1 Ϯ 1.0 CD28Ϫ/Ϫ 1/13 (7.7%) 1.0 100 0/6 (0%) 10 1/6 (16.7%) 1.0

a Mean Ϯ SEM of the maximal scores of arthritic mice. The Journal of Immunology 207

Table III. Adoptive transfer of arthritis in CD28Ϫ/Ϫ mice by LN cells CD28-deficient mice and mice injected with CTLA4-Ig exhibit from CII-immunized DBA/1 mice decreased production of Th2-induced IgG1 Abs (6, 17). Th2 responses have generally been reported to be severely impaired in Cell Cell Incidence Day of Arthritis CD28-deficient or CTLA4-Ig-treated mice, consistent with the Expt. Population Number of Disease Onset Index demonstration that CD28 promotes Th2 cytokine production (35). 1 LN cells 6 ϫ 107 2/3 6, 6 8, 8 However, in contrast, a recent study showed unimpaired Th2 re- 2 Purified T cells 3 ϫ 107 1/3 15 2 sponses in a different strain of CD28-deficient mice, suggesting ϫ 7 3 LN cells 5 10 2/3 7, 9 4, 3 that the effect of blocking CD28 in the Th1/Th2 balance varies depending on the genetic background of the mouse and the type of immunogen (36). Our results show that the IgG1 anti-CII Ab level (27, 28). The present study of CD28-deficient mice suggests an was lower than that of IgG2a in CD28Ϫ/Ϫ DBA/1 mice, whereas indispensable role for CD28 in triggering CIA in which CD28- both subclasses were at a similar level in CD28ϩ/Ϫ DBA/1 mice. mediated activation cannot be substituted for by other costimula- We measured Th2 cytokine production of LN cells in response to tory molecules. Yashiro et al. (29) recently showed the differences CII but failed to detect IL-4 or IL-10 in response to CII in between CD28 and other costimulatory molecules in the capacity CD28Ϫ/Ϫ as well as CD28ϩ/Ϫ mice. In addition, there is no dif- to activate naive T cells. Although CD28 and other costimulatory ferences in the expression of IFN-␥ mRNAs in the LNs from molecules, including CD2, CD5, CD9, CD11a, or CD44, induced CD28ϩ/Ϫ and CD28Ϫ/Ϫ mice, and only slight reduction of IL-4 T cell proliferation, only CD28 induced IL-2 production and pre- expression was observed in limited number of CD28Ϫ/Ϫ mice. In vented apoptosis after activation. These results support the nonre- a previous study, IL-4, IL-10, and IFN-␥ expressions have been Downloaded from dundant function of CD28 in the induction and survival of Ag- observed by immunization with CFA alone (33), which suggests specific T cells after active immunization. macrophages activated by adjuvant may induce those cytokines. In contrast to our results, it has been shown that CD28-deficient Furthermore, these cytokine expressions do not necessarily indi- NOD mice develop accelerated and severe diabetes (18). The dif- cate profile of CII-reactive T cells, since CII-reactive T cells may ference may first exist in the induction of disease. CIA is induced constitute only a small population in the LNs. Our results indicate by immunization with CII, whereas NOD mice spontaneously de- that although we could not reveal an altered Th1/Th2 cytokine http://www.jimmunol.org/ velop disease. In the latter case, autoantigens might be expressed profile of T cells, CD28-deficient DBA/1 mice have mounted se- for a long time, presumably from birth, which may lead to the verely impaired Th2 responses as suggested from the IgG1/IgG2a extended exposure of autoantigens to T cells, resulting in the ac- anti-CII Ab profile. However, it is unclear whether Th1/Th2 im- tivation of Ag-specific T cells in the absence of CD28. Another balance plays any role in the resistance of CD28-deficient mice to possibility may be due to the different nature of Ag in its efficacy CIA since, as mentioned before, Th1 responses are predominant to induce T cell activation. The signal in the absence of CD28 may and play a major role in the initiation of CIA. be below the level to activate and expand T cells specific to CII, In conclusion, CD28 is an essential prerequisite for cellular and whereas it may reach the threshold to activate Ag-specific T cells humoral immunity against CII, and CD28 costimulation is required and proceed following disease process in NOD mice. for the development of CIA in DBA/1 mice. Activation of CII- by guest on September 24, 2021 We further examined the susceptibility of CD28-deficient specific T cells to initiate CIA requires costimulatory signals pro- DBA/1 mice to CIA by reinforced induction protocols. Four im- vided by CD28 which cannot be replaced by other costimulatory munizations with CII in CFA resulted in a minimal response. The molecules or repeated administration of the Ag. arthritic mouse showed IgG anti-CII Ab levels comparable to control CD28ϩ/Ϫ mice; nevertheless, the disease of the arthritic mouse Acknowledgments subsided within 10 days. These results suggest that the function of We thank M. Fujisaki for technical assistance and animal care, Y. Tsugi- CD28 in disease development is not limited to the promotion of tomi for technical help with histology, Drs. Josef M. Penninger and Hans- anti-CII Ab production by B cells but includes the triggering of the Willi Mittru¨cker for critical reading of the manuscript, and Mary Saunders inflammatory action of various cellular components, presumably for scientific editing. via cytokine production. We also tested whether a low dose of IL-12 rendered CD28-deficient mice susceptible to CIA. IL-12 References promotes Th1 type T cell responses and has been shown to poten- 1. Bretscher, P. 1992. The two-signal model of lymphocyte activation twenty-one tiate various autoimmune diseases (30). In the CIA model, it has years later. Immunol. Today 13:74. been reported that the administration of IL-12 together with CII 2. Schwartz, R. H. 1990. A cell culture model for T lymphocyte clonal anergy. plus IFA induces severe arthritis in DBA/1 mice (20, 31). Con- Science 248:1349. 3. Linsey, P. S., and J. A. Ledbetter. 1993. The role of the CD28 receptor during versely, IL-12-deficient mice are resistant to CIA (33). Further- T-cell responses to antigen. Annu. Rev. Immunol. 11:191. more, it has been reported that Th1 responses are predominant at 4. June, C. H., J. A. Bluestone, L. M. Nadler, and C. B. Thompson. 1994. The B7 and CD28 receptor families. Immunol. Today 15:321. the time of onset of arthritis and that the cytokine environment 5. Harding, F. A., J. G. McArthur, J. A. Gross, D. H. Raulet, and J. P. Allison. 1992. (high levels of IFN-␥) influences the T cells responding to CII to CD28-mediated signalling co-stimulates murine T cells and prevents induction of become type 1 T cells (33, 34). These studies indicate the impor- anergy in T-cell clones. Nature 356:607. 6. Linsley, P. S., P. M. Wallace, J. Johnson, M. G. Gibson, J. L. Greene, tance of Th1 type responses in cellular immunity against CII. J. A. Ledbetter, C. Singh, and M. A. Tepper. 1992. Immunosuppression in vivo However, CD28-deficient mice treated with IL-12 developed CIA by a soluble form of the CTLA-4 T cell activation molecule. Science 257:792. of very low incidence and minimal severity, and IgG anti-CII Ab 7. Lenschow, D. J., Y. Zeng, J. R. Thistlethwaite, A. Montag, W. Brady, M. G. Gibson, P. S. Linsley, and J. A. Bluestone. 1992. Long-term survival of levels were not significantly increased. Szeliga et al. (31) showed xenogeneic pancreatic islet grafts induced by CTLA4Ig. Science 257:789. that the enhancing effect of IL-12 on anti-CII Ab production and 8. Turka, L. A., P. S. Linsley, H. Lin, W. Brady, J. M. Leiden, R. Q. Wei, incidence of CIA was not observed in B10.Q mice, suggesting M. L. Gibson, X. G. Zheng, S. Myrdal, D. Gordon, T. Bailey, S. F. Bolling, and C. B. Thompson. 1992. T-cell activation by the CD28 ligand B7 is required for genetic factors other than MHC is responsible for this phenome- cardiac allograft rejection in vivo. Proc. Natl. Acad. Sci. USA 89:11102. non. Our data indicate that CD28-deficient mice are resistant to 9. Lin H., S. F. Bolling, P. S. Linsley, R. Q. Wei, D. Gordon, C. B. Thompson, and L. A. Turka. 1993. Long-term acceptance of major histocompatibility complex IL-12-induced enhancement of anti-CII responses and that IL-12 mismatched cardiac allografts induced by CTLA4Ig plus donor-specific transfu- cannot overcome the absence of CD28. sion. J. Exp. Med. 178:1801. 208 COLLAGEN-INDUCED ARTHRITIS IN CD28-DEFICIENT MICE

10. Wallace, P. M., J. S. Johnson, J. F. MacMaster, K. A. Kennedy, P. Gladstone, and 24. Lucas, P. J., I. Negishi, K. Nakayama, L. E. Fields, and D. Y. Loh. 1995. Naive P. S. Linsley. 1994. CTLA4Ig treatment ameliorates the lethality of murine graft- CD28-deficient T cells can initiate but not sustain an in vitro antigen-specific versus-host disease across major histocompatibility complex barriers. Transplan- immune response. J. Immunol. 154:5757. tation 58:602. 25. Biancone, L., G. Andres, H. Ahn, A. Lim, C. Dai, R. Noelle, H. Yagita, 11. Finck, B. K., P. S. Linsley, and D. Wofsy. 1994. Treatment of murine lupus with C. De Martino, and I. Stamenkovic. 1996. Distinct regulatory roles of lymphocyte CTLA4Ig. Science 265:1225. costimulatory pathways on T helper type-2 mediated autoimmune disease. J. Exp. 12. Lenschow, D. J., S. C. Ho, H. Sattar, L. Rhee, G. Gray, N. Nabavi, K. C. Herold, Med. 183:1473. and J. A. Bluestone. 1995. Differential effects of anti-B7-1 and anti-B7-2 mono- 26. Griggs, N. D., S. S. Agersborg, R. J. Noelle, J. A. Ledbetter, P. S. Linsley, and clonal antibody treatment on the development of diabetes in the nonobese dia- K. S. K. Tung. 1996. The relative contribution of the CD28 and gp39 costimu- betic mouse. J. Exp. Med. 181:1145. latory pathways in the clonal expansion and pathogenic acquisition of self-reac- 13. Cross, A. H., T. J. Girard, K. S. Giacoletto, R. J. Evans, R. M. Keeling, R. F. Lin, tive T cells. J. Exp. Med. 183:801. J. L. Trotter, and R. W. Karr. 1995. Long-term inhibition of murine experimental 27. Ranges, G. E., S. Sriram, and S. M. Cooper. 1985. Prevention of type II collagen- autoimmune encephalomyelitis using CTLA-4-Fc supports a key role for CD28 induced arthritis by in vivo treatment with anti-L3T4. J. Exp. Med. 162:1105. costimulation. J. Clin. Invest. 95:2783. 14. Arima, T., A. Rehman, W. F. Hickey, and M. W. Flye. 1996. Inhibition by 28. Hom, J. T., L. D. Butler, P. E. Riedl, and A. M. Bendele. 1988. The progression CTLA4Ig of experimental allergic encephalomyelitis. J. Immunol. 156:4916. of the inflammation in established collagen-induced arthritis can be altered by 15. Knoerzer, D. B., R. W. Karr, B. D. Schwartz, and L. J. Mengle-Gaw. 1995. treatments with immunological or pharmacological agents which inhibit T cell Collagen-induced arthritis in the BB rat: prevention of diseases by treatment with activities. Eur. J. Immunol. 18:881. CTLA-4-Ig. J. Clin. Invest. 96:987. 29. Yashiro, Y., X.-G. Tai, K. Toyo-oka, C.-S. Park, R. Abe, T. Hamaoka, 16. Webb, L. M. C., M. J. Walmsley, and M. Feldmann. 1996. Prevention and ame- M. Kobayashi, S. Neben, and H. Fujiwara. 1998. A fundamental difference in the lioration of collagen-induced arthritis by blockade of the CD28 co-stimulatory capacity to induce proliferation of naive T cells between CD28 and other co- pathway: requirement for both B7-1 and B7-2. Eur. J. Immunol. 26:2320. stimulatory molecules. Eur. J. Immunol. 28:926. 17. Shahinian, A., K. Pfeffer, K. P. Lee, T. M. Kundig, K. Kishihara, A. Wakeham, 30. Trembleau, S., T. Germann, M. K. Gately, and L. Adorini. 1995. The role of K. Kawai, P. S. Ohashi, C. B. Thompson, and T. W. Mak. 1993. Differential T IL-12 in the induction of organ-specific autoimmune diseases. Immunol. Today cell costimulatory requirements in CD28-deficient mice. Science 261:609. 16:383. Downloaded from 18. Lenschow, D. J., K. C. Herold, L. Rhee, B. Patel, A. Koons, H.-Y. Quin, 31. Szeliga, J., H. Hess, E. Rude, E. Schmitt, and T. Germann. 1996. IL-12 promotes E. Fuchs, B. Singh, C. B. Thompson, and J. A. Bluestone. 1996. CD28/B7 reg- cellular but not humoral type II collagen-specific Th1-type responses in C57BL/6 ulation of Th1 and Th2 subsets in the development of autoimmune diabetes. and B10.Q mice and fails to induce arthritis. Int. Immunol. 8:1221. Immunity 5:285. 32. McIntyre, K. W., D. J. Shuster, K. M. Gillooly, R. R. Warrier, S. E. Connaughton, 19. Bachmaier, K., C. Pummerer, A. Shahinian, J. Ionescu, N. Neu, T. W. Mak, and L. B. Hall, L. H. Arp, M. K. Gately, and J. Magram. 1996. Reduced incidence and J. M. Penninger. 1996. Induction of autoimmunity in the absence of CD28 co- severity of collagen-induced arthritis in interleukin-12-deficient mice. Eur. J. Im- stimulation. J. Immunol. 157:1752. munol. 26:2933. 20. Germann, T., J. Szeliga, H. Hess, S. Storkel, F. J. Podlaski, M. K. Gately, 33. Stasiuk, L. M., O. Abehsira-Amar, and C. Fournier. 1996. Collagen-induced ar- E. Schmitt, and E. Rude. 1995. Administration of interleukin 12 in combination http://www.jimmunol.org/ thritis in DBA/1 mice: cytokine gene activation following immunization with with type II collagen induces arthritis in DBA/1 mice. Proc. Natl. Acad. Sci. USA type II collagen. Cell. Immunol. 173:269. 92:4823. 21. Tada, Y., A. Ho, D.-R. Koh, and T. W. Mak. 1996. Collagen-induced arthritis in 34. Mauri, C., R. O. Williams, M. Walmsley, and M. Feldmann. 1996. Relationship CD4- or CD8-deficient mice: CD8ϩ T cells play a role in initiation and regulate between Th1/Th2 cytokine patterns and the arthritogenic response in collagen- recovery phase of collagen-induced arthritis. J. Immunol. 156:4520. induced arthritis. Eur. J. Immunol. 26:1511. 22. Tada, Y., A. Ho, T. Matsuyama, and T. W. Mak. 1997. Reduced incidence and 35. Rulifson, I. C., A. I. Sperling, P. E. Fields, F. W. Fitch, and J. A. Bluestone. 1997. severity of antigen-induced autoimmune diseases in mice lacking interferon reg- CD28 costimulation promotes the production of Th2 cytokines. J. Immunol. 158: ulatory factor-1. J. Exp. Med. 185:231. 658. 23. Kundig, T. M., A. Shahinian, K. Kawai, H.-W. Mittrucker, E. Sebzda, 36. Brown, D. R., J. M. Green, N. H. Moskowitz, M. Davis, C. B. Thompson, and M. F. Bachmann, T. W. Mak, and P. S. Ohashi. 1996. Duration of TCR stimu- S. L. Reiner. 1996. Limited role of CD28-mediated signals in T helper subset lation determines costimulatory requirement of T cells. Immunity 5:1. differentiation. J. Exp. Med. 184:803. by guest on September 24, 2021