Dipeptidyl peptidase IV on activated T cells as a target molecule for therapy of rheumatoid arthritis

Abstract

The extracellular domain of the T cell co-stimulatory molecule CD26 possesses dipeptidyl peptidase IV (DP IV) enzyme activity. Activated T cells are known to increase expression of cell surface DP IV and some specific inhibitors of this enzyme have been reported to suppress T cell function. Previously we have identified a DP IV inhibitor, designated TMC-2, found in culture supernatant of Aspergillus oryzae. Administration of TMC-2 to rats with adjuvant arthritis caused marked suppression of paw swelling. To elucidate the mechanism of TMC-2 antiarthritic activity, we have studied its effects on T cell function. Here we show that TMC-2 inhibited DP IV activity of CD26 immunoprecipitated from T cell lysates, and also inhibited proliferative responses of T cells to specific antigen or anti-CD3 antibody. Suppression of IL-2 production was demonstrated at both the mRNA and protein levels. TMC-2 did not alter the PTPase activity of pure CD45, but when this molecule was co-precipitated from T cell lysates together with associated CD26, its PTPase was virtually completely abolished by TMC-2. These results suggest that modulation of CD45 PTPase activity might be responsible for functional suppression of T cells by TMC-2. Because the effects of TMC-2 on T cells were reversible and it was not toxic at the concentrations used, TMC-2 may be a candidate novel therapeutic agent for rheumatoid arthritis.

INTRODUCTION

Dipeptidyl peptidase IV (DP IV, EC 3·4.14·5) is a widely distributed ectoenzyme in mammalian tissues such as the liver, kidney, small intestine, salivary gland and blood cells. DP IV preferentially cleaves amino-terminal dipeptides from polypeptides with proline or alanine at the penultimate position, including several hormones, neuropeptides, cytokines and chemokines [1,2]. In the immune system, the extracellular carboxy-terminal domain of CD26, one of the accessory molecules of helper T cells, has DP IV activity, and mediates a co-stimulatory effect on T cell activation via the CD3 and CD2 pathways [3]. In addition to CD26, however, other cell surface molecules that harbour DP IV activity have been identified on T cells, such as DPP IV-β[4] and DPPT-L [5], and possibly another molecule on B lineage cells [6]. The soluble form of DP IV in serum, thought to be secreted from activated T cells, has immunostimulatory activity and, at least in part, consists of DPPT-L [7]. Despite numerous studies, both the natural ligands of CD26 and substrates of DP IV have been elusive.

Studies using specific inhibitors have revealed an involvement of DP IV in T cell function. Peptides containing the α-amino boronic acid analogue of proline (boroPro) as the carboxy-terminal residue, Ala-boroPro and Pro-boroPro, inhibit DP IV with Ki values in the nanomolar range, and also inhibit antigen-induced lymphocyte proliferation and IL-2 production in cultures of murine T cells [8]. Similarly, Pro-boroPro inhibited tetanus toxoid-stimulated human T cell responses, which was overcome by the addition of IL-2, suggesting that inhibition of DP IV resulted in a state of anergy, probably by interfering with delivery or amplification of a signal necessary for IL-2 production [9]. We have previously observed that production of antigen-specific IgG antibodies in normal BALB/c mice was suppressed by administration of these DP IV inhibitors, demonstrating that DP IV plays a role in immune responses in vivo[10]. Other investigators reported an increase of TGF-β1 production caused by a DP IV inhibitor Lys[Z(NO₂)]-pyrrolidide, resulting in normalization of the Th1-Th2 balance and amelioration of autoimmune encephalomyelitis in a mouse model [11].

In an attempt to apply these findings to developing a new mode of therapy of inflammatory disorders we have recently administered a novel DP IV inhibitor, TMC-2, to rats with adjuvant-induced arthritis and observed significant reduction of swelling of the paws, splenomegaly and weight loss, as well as decreased levels of serum mucoprotein [12]. TMC-2 is an antibiotic produced by Aspergillus oryzae and purified from its culture supernatant. It inhibits DP IV, but not other proteases tested, with a Ki value of ∼5 µm in a non-competitive manner [13]. In this communication, we analyse the mechanisms of the antirheumatic effects of TMC-2.

MATERIALS AND METHODS

DP IV enzyme activity and a DP IV inhibitor

Samples were incubated at 37°C with 3 mm of the substrate Gly-Pro-p-nitroanilide (Sigma, St Louis, MO, USA) in 25 mm HEPES buffer (pH 7·8) containing 140 mm NaCl and the O.D. at 405 nm was monitored. One unit of DP IV activity was defined as the amount of enzyme catalysing the formation of 1 µmol/min of p-nitroaniline at 37°C. A DP IV inhibitor, TMC-2, was purified from culture supernatant of A. oryzae A374 as described previously and aliquots of the sterile stock solution were kept at −30°C until use [13]. Human serum was obtained from healthy donors after giving informed consent.

Proliferative responses of T cells

Murine splenocytes were prepared by hypotonic lysis to remove red blood cells from BALB/c mice immunized with 20 µg of purified protein tuberculin derivative (PPD; Difco, Detroit, MI, USA) 3 months previously. Human peripheral blood mononuclear cells (PBMC) were obtained by density gradient centrifugation from healthy donors after giving informed consent. Murine splenocytes and human PBMC were cultured at 5 × 10⁵ cells/ml in complete RPMI-1640 medium containing 10% FCS (GIBCO, Rockville, MD, USA) for 5 days. At the beginning of the culture, different concentrations of TMC-2 were added and 15 min later the murine cells and human cells were stimulated with 50 µg/ml and 20 µg/ml PPD, respectively.

In another experiment, human T cells were enriched from PBMC by passing over a nylon wool column. This fraction of the cell preparations contained more than 90% T cells as assessed by flow cytometry with anti-CD2 monoclonal antibody (Becton Dickinson, Franklin Lakes, NJ, USA). These cells were incubated at 1 × 10⁶ cells/ml in serum-free medium with different concentrations of TMC-2 for 15 min in round-bottomed tubes. Then FCS (final 10%) was added and the cells were transferred into 96-well plates coated with the 2 µg/ml anti-CD3 monoclonal antibody UCHT1 (Ancell, Bayport, MN, USA) and cultured for 4 days. In both murine and human experiments, incorporation of [³H]TdR during the last 16 h of the culture was measured by a scintillation counter.

Measurement of secreted IL-2 and expression of IL-2 mRNA

Human peripheral blood T cells prepared as above (1 × 10⁶/ml) were incubated in serum-free medium with or without TMC-2 for 15 min, then FCS (final 10%) was added and the cells were transferred into anti-CD3-coated 96-well plates. The supernatant was harvested after 24, 48 or 72 h, and IL-2 concentrations were measured using an IL-2 enzyme immunoassay kit (Immunotech, Marseilles, France).

Total RNA from 1 × 10⁷ human peripheral blood T cells cultured as above for 4, 24 or 48 h on anti-CD3-coated plates was isolated using the RNeasy kit (Qiagen, Tokyo, Japan). After the first isolation, contaminating DNA was removed by DNase I (Qiagen) digestion and the RNA was further purified using RNeasy. Reverse transcription and PCR were carried out sequentially using the One Step RT-PCR kit (Qiagen). One hundred ng of total RNA was reverse transcribed for 30 min at 50°C, and cDNA was amplified with oligonucleotide primers specific for the human IL-2 gene (sense; 5′-ATGTACAGGATGCAACTCCTGTCTT-3′ and antisense; 5′-GTCAGTGTTGAGATGATGCTTTGAC-3′), and the glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH; sense; 5′-CCACCCATGGCAAATTCCATGGCA-3′ and antisense; 5′-TCTAGACGGCAGGTCAGGTCCACC-3′). The products were analysed by 2% agarose gel electrophoresis. To exclude the contamination by genomic DNA, negative control tubes were set up with inactivated reverse transcriptase by heating at 95°C at the beginning of each RT-PCR reaction.

Immunoprecipitation of CD26 and CD45

Cell lysates were prepared from 1 × 10⁷ cells of the CCRF-HSB-2 human T cell leukaemia line (obtained from the Institute of Physical and Chemical Research, Tsukuba, Japan) in Triton lysis buffer (0·5% Triton X-100, 150 mm NaCl, 5 mm MgCl₂, 5 mm 2-mercaptoethanol, 100 U/ml aprotinin, 0·2 mm PMSF, 20 mm Tris, pH 7·6) or in NP-40 lysis buffer (1% NP-40, 150 mm NaCl, 1 mm PMSF, 5 mm EDTA, 10 mm Tris, pH 7·4). Anti-CD26 monoclonal antibody Ta1 (final 80 µg/ml, Coulter Immunology, Hialeah, FL, USA) or anti-CD45 monoclonal antibody C11 (final 80 µg/ml, Ancell) was added to the lysates with rabbit antimouse IgG serum and Protein A-Sepharose (Amersham Pharmacia Biotec, Tokyo, Japan), and incubated with gentle rotation for 2 h at 4°C. After centrifugation, the precipitates were washed three times with the lysis buffer and used for Western blotting or enzyme assays. For Western blotting, the proteins were solubilized in SDS-PAGE sample buffer (1% SDS, 5% 2-mercaptoethanol, 5% glycerol, 30 mm Tris, pH 6·8), heated for 5 min at 98°C, and separated by 6% SDS-PAGE. After transfer to PVDF membranes (Millipore, Tokyo, Japan), the proteins were probed by Ta1 and C11, and visualized by peroxidase-conjugated secondary antibodies using the ECL system (Amersham Pharmacia Biotec). To determine the enzyme activity, the precipitates were suspended in HEPES buffer as described above for DP IV activity or in Tris-buffered serine (25 mm Tris, 140 mm NaCl, pH 7·4) for PTPase activity.

Phosphatase activity of CD45 and calcineurin

CD45 tyrosine phosphatase (PTPase) activity was measured using a kit obtained from Biomol (Plymouth Meeting, PA, USA), which consists of recombinant human CD45, negative regulatory site peptide of an Src family kinase as a substrate and RWJ-60475 as a PTPase inhibitor. Enzyme activity was measured by quantification of the released phosphate by a modified malachite green assay according to the manufacture's instructions, monitoring O.D. at 620 nm. One unit of CD45 PTPase activity was defined as the amount of enzyme catalysing the formation of 1 nmol/min free phosphate at 25°C. Phosphatase activity of recombinant calcineurin was measured using a kit obtained from Calbiochem (San Diego, CA, USA) according to the manufacturer's protocol.

RESULTS

Inhibition of serum soluble DP IV and CD26 DP IV

Before beginning this investigation of the mechanisms of the anti-arthritic effect of TMC-2, we first confirmed its inhibitory effect on DP IV activity. Human serum obtained from healthy volunteers was incubated with different concentrations of TMC-2 for 15 min at 37°C and enzyme activity of soluble DP IV was measured using Gly-Pro-p-nitroanilide as a substrate. TMC-2 inhibited DP IV activity with an IC₅₀ value of ∼3 µm, as shown in Fig. 1a. To determine whether TMC-2 also inhibits the DP IV activity of CD26, immunoprecipitation was carried out using cell lysates from the human T cell line CCRF-HSB-2 and anti-CD26 monoclonal antibody Ta1. DP IV activity was detected in the precipitated fraction and it was inhibited by TMC-2 in a dose-dependent manner, again with an IC₅₀ of less than 10 µm (Fig. 1b). The DP IV activity was not recovered at all when the samples were kept for 24 h at 37°C with TMC-2 (data not shown), indicating that TMC-2 forms stable complexes with both the soluble form of DP IV in serum and with T cell surface CD26 DP IV.

Fig. 1.

Inhibition of soluble DP IV in serum and CD26 DP IV on T cells by TMC-2. Healthy human serum (a) or CD26 immunoprecipitated by anti-CD26 monoclonal antibody Ta1 from cell lysate of CCRF-HSB-2 human T cells (b) was incubated for 15 min with TMC-2, and DP IV activity was measured at 37°C. In (b), the vertical axis expresses enzyme activity of 300 µl suspension containing Protein A-Sepharose beads bound with CD26 derived from 8·5 × 10⁶ cells. Control: antimouse IgG and Protein A-Sepharose were added without Ta1 to the immunoprecipitate.

Suppression of proliferation and IL-2 production of T cells

To examine the functional effect of the binding of TMC-2 to T cell surface DP IV, spleen cells from BALB/c mice primed with PPD were stimulated in vitro with PPD in the presence or absence of TMC-2. As shown in Fig. 2a, the proliferative response of murine spleen cells to PPD was markedly suppressed by TMC-2 in a dose-dependent manner. Similarly, the proliferative response of PBMC, obtained from healthy donors who were positive for PPD skin test to PPD, was suppressed by TMC-2 (Fig. 2b). At a concentration of 100 µm, more than 85% of [³H]TdR incorporation by murine and human T cells was suppressed. On the other hand, suppression of human peripheral blood T cells stimulated by immobilized anti-CD3 antibody required a higher concentration of TMC-2 compared to the antigen-specific responses (Fig. 2c). In this experiment, viability of the cells at the end of the culture was estimated by a dye-exclusion test, and was found to be greater than 98% even at a TMC-2 concentration of 300 µm.

Fig. 2.

Inhibition of proliferative responses by TMC-2. Spleen cells prepared from BALB/c mice immunized with PPD (a), or human PBMC (b) were incubated with or without TMC-2 for 15 min, PPD was added, and cells were cultured for 5 d. (c) Human peripheral blood T cells were incubated with or without TMC-2 for 15 min and transferred into anti-CD3-coated plates, and then further cultured for 4 d. In these experiments, [³H]TdR incorporation during the last 16 h of the culture was measured. Data are expressed as mean ± s.d. of triplicate samples. In (c), viability of the cells at the end of the culture was also determined by trypan blue dye-exclusion (•).

To explore whether the suppressive effect of TMC-2 on T cell proliferative responses is accompanied by suppression of cytokine production, we measured IL-2 concentrations by ELISA in the culture supernatants of human peripheral blood T cells stimulated with immobilized anti-CD3, in the presence or absence of TMC-2. It was found that 300 µm TMC-2 significantly suppressed IL-2 production during 24 or 48 h (Fig. 3a). When the cells were cultured for 72 h, however, this effect became insignificant compared to cultures without TMC-2. Results of RT-PCR assays also showed significant suppression of IL-2 mRNA expression by TMC-2 at 4–48 h, with no reduction of the mRNA level of the GAPDH house-keeping gene (Fig. 3b).

Fig. 3.

Suppression of IL-2 production by TMC-2. (a) Human peripheral blood T cells were incubated for 15 min with or without 300 µm TMC-2, transferred into anti-CD3-coated plates and cultured for 24, 48 or 72 h. IL-2 concentration in the culture supernatants was measured by ELISA. (b) Human peripheral blood T cells were treated as above, cultured for 4, 24 or 48 h, and total RNA was extracted and expression of IL-2 mRNA was estimated by RT-PCR. , TMC-2(−); ▪, TMC-2(+).

Effect of TMC-2 on the upstream component of the T cell activation pathway

Because CD26 has only six amino acids in the cytoplasmic tail, it has been speculated that other molecules are necessary for CD26-mediated signal transduction [1,2]. One of the candidates for this signal transduction molecule is CD45, which is a major transmembrane PTPase on the surface of activated T cells and is thought to be associated with CD26 [14]. CD45 PTPase is believed to dephosphorylate and activate Src family kinases, and CD45-deficient cells do not exhibit normal intracellular activation events [15]. Therefore, we studied whether TMC-2 affects CD45 PTPase activity in solubilized T cell membrane fractions. Cell lysates were prepared from CCRF-HSB-2 T cells and incubated with or without the DP IV inhibitor. Fifty to 100 µm of TMC-2 inhibited DP IV activity in the lysate by more than 77%, and interestingly also inhibited PTPase activity by ∼50% (Fig. 4a). To confirm that this PTPase activity was derived from CD45, immunoprecipitation was carried out using an anti-CD26 or an anti-CD45 monoclonal antibody. As expected from a previous report [14], anti-CD26 antibody could co-precipitate CD45 and vice versa (Fig. 4b, insert). Furthermore, not only DP IV activity but also PTPase activity in these precipitates was suppressed virtually completely by 300 µm TMC-2 (Fig. 4b). This was not a result of non-specific reactions of TMC-2, because TMC-2 did not directly suppress PTPase activity of recombinant CD45 alone (Fig. 5).

Fig. 4.

Effect of TMC-2 on CD45 PTPase activity. (a) Cell lysate prepared from CCRF-HSB-2 T cells was incubated with 300 µm TMC-2 for 15 min, and DP IV (▪) and PTPase () activity was measured. 100% activity of DP IV and PTPase corresponds to 2·5 U/l and 5700 U/l, respectively. (b) From the T cell lysate, both CD26 and CD45 were co-precipitated by either anti-CD26 antibody (left lane) or anti-CD45 antibody (right lane) as shown in the insert, indicating association of these molecules in the cell membrane. Immunoprecipitate either by anti-CD26 antibody (▪) or anti-CD45 antibody () was incubated with 300 µm TMC-2 for 15 min and PTPase activity was measured. 100% of CD45 PTPase activity precipitated by anti-CD26 antibody and anti-CD45 antibody corresponds to 2000 U/l and 2300 U/l, respectively.

Fig. 5.

No direct inhibition of CD45 PTPase by TMC-2. Human recombinant CD45 PTPase and a substrate peptide were incubated for 60 min at 25°C with various concentrations of a PTPase inhibitor RWJ-60475, or a DP IV inhibitor TMC-2. The released phosphate was quantified by a modified malachite green assay.

To confirm specificity, the effect of TMC-2 on calcineurin, which is another phosphatase relevant to T cell activation, was also tested. As a result, 100–300 µm of TMC-2 did not directly suppress enzyme activity of recombinant calcineurin (data not shown).

Reversibility of the TMC-2-induced functional arrest of T cells

Viability of the cells remained nearly 100% and mRNA levels of the control GAPDH were unaffected at concentrations of TMC-2, which strongly suppressed T cell function. To confirm further that the effects of TMC-2 on T cell function are not due to cell death, the agent was washed out after CCRF-HSB-2 cells had been treated for 15 min, and these cells were then cultured in TMC-2-free medium for 16–60 h. When the cells were treated and cultured at a density of 1 × 10⁶ cells/ml, [³H]TdR incorporation during the first 16 h of the culture was ∼20% suppressed by the short-term treatment with TMC-2, but had recovered completely by 40 h of culture. When the cells were treated at lower cell densities and cultured at 1 × 10⁶ cells/ml, suppressive effects of the short-term treatment with TMC-2 were more significant; however, DNA synthesis of the cells recovered completely in all cases by 40 h (Fig. 6). These results illustrate two important findings: first, the functional suppression of the T cells by TMC-2 is reversible, and secondly, the effect depends on the relative concentration of the TMC-2 to the cell density.

Fig. 6.

Reversibility of the suppressive effect of TMC-2 on a T cell proliferative response. CCRF-HSB-2 human T cells at the density of 1 × 10⁵, 5 × 10⁵ or 1 × 10⁶ cells/ml as indicated in the graphs were incubated with TMC-2 for 15 min in serum-free medium, washed with fresh serum-free medium, and cultured in TMC-2-free complete medium for 16, 40 or 64 h at the density of 1 × 10⁶ cells/ml. [³H]TdR incorporation during the last 16 h of the culture was measured. Data are expressed as mean ± s.d. of triplicate samples. *P < 0·05, **P < 0·01, P-values indicate significant differences compared to 0 µm. ▪, 15 h; , 40 h; , 64 h.

DISCUSSION

We have shown previously that DP IV activity both in the serum and at the lymphocyte surface changes in parallel after immunization of normal BALB/c mice with BSA and adjuvant, which suggested that serum DP IV is secreted from activated T cells [10]. Duke-Cohan et al. [5] subsequently demonstrated that the 175 kDa form of soluble DP IV in human serum is secreted by activated T cells but is distinct from 105 kDa CD26. In the present study, we first observed that both soluble DP IV in human serum and CD26 DP IV immunoprecipitated from human T cell lysates were inhibited by TMC-2 with a similar IC₅₀ value. Most of the experiments in this study were carried out using human T cells because of the availability of monoclonal antibodies; however, we have confirmed that TMC-2 also inhibited murine DP IV.

TMC-2 exhibits a potent anti-arthritic effect on adjuvant-induced rat arthritis [12]. Inhibition of DP IV activity in the serum possibly contributes to this anti-arthritic effect, because soluble DP IV has been shown to possess immunostimulatory activity [5]. However, we have focused here on the effect of TMC-2 on T cells. Because activated T cells strongly expressing surface DP IV infiltrate into the synovial tissue and play a role in arthritis [16], and several DP IV inhibitors have been reported to inhibit T cell activation [8,17,18], it may be hypothesized that one of the mechanisms of the anti-arthritic effect of TMC-2 is related to T cell function. In accordance with this hypothesis, we found that TMC-2 suppressed proliferative responses of both murine and human T cells stimulated by specific antigen or anti-CD3 antibody. We noticed, however, that the concentration of TMC-2 required for>50% suppression of [³H]TdR incorporation varied between experiments, and it seems likely that it is not the absolute TMC-2 concentration in the medium but the concentration of TMC-2 relative to cell density which is crucial. This is shown clearly in Fig. 6, in which the suppressive effect of 300 µm TMC-2 was related inversely to cell density. Thus, in the experiments shown in Fig. 2, lower concentrations of TMC-2 were sufficient for significant inhibition of PPD-stimulated T cell responses than required for an anti-CD3-stimulated T cell response, mainly because in the former, the responding cell fraction contains a much lower number of reactive T cells strongly expressing DP IV than the latter where immobilized anti-CD3 antibody stimulates a far larger proportion of the T cells.

Suppression of T cell proliferative responses by TMC-2 does not result from cell death caused by non-specific cytotoxicity, because the viability of the cells remained high at those concentrations of TMC-2 which significantly inhibited [³H]TdR incorporation (Fig. 2c), and the mRNA level of a housekeeping gene was not affected (Fig. 3b). No staining of human PBMC with annexin V after 20 h incubation with 0–300 µm of TMC-2 was observed (data not shown). In addition, the suppressive effect was reversible in the experiments in which we washed the cells after brief treatment with TMC-2 (Fig. 6). In our previous study of daily administration of TMC-2 to arthritic rats, physical activity and body weight of the animals improved significantly in the TMC-2-injected group compared to the untreated control [12]. These results indicate that TMC-2 does not possess severe cytotoxicity at the effective treatment dose, and could therefore be a candidate for a novel therapeutic agent in rheumatoid arthritis. A clinical trial for treatment of diabetes mellitus using a different DP IV inhibitor has recently been undertaken, in which the inhibitor suppresses cleavage of glucagon-like peptide by DP IV [19]. This study encourages us to pursue the possibility of clinical application of DP IV inhibitors to various diseases in which DP IV is involved.

Regarding the molecular mechanism of the anti-arthritic effect of TMC-2, it is unlikely that inhibition of DP IV enzyme activity per se is critical for T cell function, because animals with mutated DP IV that lacks enzyme activity still show normal immune responses [20; our unpublished observations]. Rather, the binding of an inhibitor to the specific enzyme may affect functions of other molecules associating with CD26. CD26 has only six amino acids in the cytoplasmic region and it requires associating molecules to transfer extracellular events into the intracellular signalling cascades [1,2]. One of the molecules that has been suggested to be associated with CD26 is CD45, which is a major membrane PTPase of activated T cells [15]. Consistent with a previous report, we observed CD26 in a fraction of T cell lysate immunoprecipitated by anti-CD45 antibody and reciprocally, CD45 was precipitated by anti-CD26 antibody. A possible explanation invokes a conformational change of the CD26 molecule bound by TMC-2 suppressing indirectly the CD45 PTPase activity that dephosphorylates the negative regulatory site of Lck kinase, thereby resulting in sequential suppression of tyrosine phosphorylation of ζ chains and downstream molecules. Our results endorsed this hypothesis by showing that CD45 PTPase activity was suppressed by TMC-2 only when it was in the presence of CD26. However, complete elucidation of the relationship between these observations and downstream events awaits further study.