Introduction
CD26 is a 110-kDa type II membrane-bound glycoprotein with dipeptidyl peptidase IV (DPPIV) activity in its extra cellular domain[
1‐
3]. CD26 is composed of 766 amino acids (AAs), and is anchored to the lipid bilayer by a single hydrophobic segment at residues 7–28. The cytoplasmic tail of CD26 is composed of only 6 amino acid residues at the N-terminal end (AA 1–6) without a common signaling motif structure. The predominant part of CD26 consists of the extra cellular domain (AA 29–766) divided into three regions, a glycosylated region, a cysteine-rich region and a C-terminal DPPIV catalytic region[
4,
5]. DPPIV belongs to the serine protease family, able to cleave dipeptides from polypeptides with N-terminal penultimate proline or alanine, and regulates the activities of a number of cytokines and chemokines[
3]. CD26 is a multifunctional molecule associated with various proteins such as adenosine deaminase (ADA), caveolin-1, CXCR4, collagen, and fibronectin, and is expressed on various cell types, including epithelial cells (kidney proximal tubules, bile duct, prostate and intestine), endothelial cells as well as T lymphocytes[
4‐
6]. The function of CD26 is dependent on cell types and the microenvironment, which influence its multiple biological roles[
4‐
7].
In addition to being a marker of T cell activation, CD26 is associated with T cell signal transduction processes as a costimulatory molecule[
4]. While CD26 expression is increased following activation of resting T cells, CD4
+CD26
high
T cells respond maximally to recall antigens such as tetanus toxoid[
8]. Moreover, crosslinking of CD26 and CD3 with solid-phase immobilized monoclonal antibodies (mAbs) can induce T cell costimulation and IL-2 production by CD26
+ T cells[
4]. Furthermore, high CD26 cell surface expression in CD4
+ T cells is correlated with the production of T
H1-type cytokines and high migratory activity[
4]. Taking into account the data that effector T cells in inflamed lesions express high levels of CD26, it is conceivable that CD4
+CD26
+ T cells play an important role in the inflammatory process[
5,
9,
10]. We have recently found that cytotoxic activity of CD8
+ T cells is also regulated via CD26-mediated costimulation[
11]. More recently, we have shown that humanized anti-CD26 mAb appears to be a promising novel therapy for the clinical control of graft-versus-host disease (GVHD) in a xenogeneic GVHD murine model[
12]. CD26 is also expressed on various tumors such as malignant mesothelioma, renal carcinoma, colon cancer, hepatocellular carcinoma, lung cancer, prostate cancer, gastrointestinal stromal tumor (GIST), thyroid cancer, T-lymphoblastic lymphoma and T-acute lymphoblastic leukemia[
13]. We have shown that administration of anti-CD26 mAb resulted in both in vitro and in vivo inhibition of tumor cell growth, migration and invasion, and enhanced survival of mouse xenograft models inoculated with T-lymphoma, renal cell carcinoma or malignant mesothelioma[
14‐
16]. Based on these findings, we have focused on CD26 as a novel therapeutic target for various tumors and immune disorders, and have developed a humanized anti-CD26 mAb, YS110, which is being investigated currently in a phase I clinical trial for patients with CD26-expressing tumors, including malignant mesothelioma[
17].
The development of companion diagnostic agents to be used in conjunction with the appropriate therapeutic modalities is essential to maximize therapeutic effectiveness while minimizing associated toxicities. Detection of tumor CD26 expression is critical to determining potential eligibility for treatment with humanized anti-CD26 mAb, and it is also important to determine whether CD26 expression on tumors or lymphocytes is affected by anti-CD26 mAb therapy. Immunohistochemical staining of CD26 with the many anti-CD26 mAbs previously developed in our laboratory[
18] did not reveal an anti-CD26 mAb that can clearly detect the denatured CD26 molecule in formalin-fixed paraffin-embedded tissues. Meanwhile, testing of several commercially available anti-CD26 mAbs designated as research reagents for immunohistochemical staining, and a mAb purchased from MBL indicated that these mAbs could stain the denatured CD26 in formalin-fixed tissues, but not with sufficient clarity. On the other hand, our testing of commercially available anti-CD26 polyclonal antibodies (pAbs), and a pAb purchased from R&D Systems showed that these reagents exhibited reliable staining pattern and intensity[
19]. However, the disadvantage of pAbs is the potential lot-to-lot variability in staining pattern and intensity, which makes pAbs not the ideal reagents for diagnostic testing of patient specimens in the clinical setting, where consistency and uniformity are required.
In the present study, by immunizing mice with CD26 protein denatured in urea buffer, we have succeeded in developing novel anti-human CD26 mAbs that can be used as companion diagnostic reagents suitable for immunohistochemical staining of CD26 in formalin-fixed tissue sections with reliable clarity and intensity. In addition, since some of these mAbs display no cross-reactivity with the therapeutic humanized anti-CD26 mAb, they may be suitable for assays analyzing CD26 expression during or following treatment with the humanized anti-CD26 mAb.
Materials and methods
Animals
Female BALB/c mice were purchased from CLEA Japan (Tokyo, Japan) and housed in a specific pathogen-free facility in micro-isolator cages. Animal experiments were conducted following protocols approved by the Animal Care and Use Committee at Juntendo University.
Antibodies
To determine the epitope of the newly developed mouse anti-human CD26 mAbs, murine anti-human CD26 mAbs (clone 4G8, 1F7, 14D10, 5F8, 16D4B or 9C11) which have been already developed in our laboratory were used[
18]. We have previously shown that these mAbs are divided into 5 separate groups by their epitopes, 4G8 recognizing the 1-247th AAs region of CD26, 1F7 and 14D10 recognizing the 248-358th AAs region of CD26, 5F8 recognizing the 359-449th (closer to the 359th) AAs region of CD26, 16D4B recognizing the 450-577th AAs region of CD26, and 9C11 recognizing the 359-653th (but different from 5F8 or 16D4B) AAs region of CD26. The humanized anti-CD26 mAb (YS110) was generated by utilizing the complementarity determining regions of the murine anti-human CD26 mAb 14D10[
18], and generously provided by Y's Therapeutics (Tokyo, Japan). To compare the staining pattern and intensity of human CD26 on formalin-fixed tissue sections, we used two commercial anti-human CD26 Abs available for CD26 detection by immunohistochemistry. One is the culture supernatant form of a mouse anti-human CD26 mAb (clone 44–4) purchased from MBL (Nagoya, Japan), and the other is a purified goat anti-human CD26 pAb purchased from R&D Systems (Minneapolis, MN). Human polyclonal IgG (venilon-I) was purchased from Alfresa Corporation (Tokyo, Japan), and mouse IgG
1 isotype control (clone MG1-45) was purchased from BioLegend (San Diego, CA). YS110, control human IgG, 4G8, 1F7, 5F8, 16D4B, 9C11, purified clone 18, clone 19, and mouse IgG
1 isotype control were labeled using an Alexa Fluor 647 Monoclonal Antibody Labeling Kit (Molecular Probes, Eugene, OR) according to the manufacturer's instructions.
cDNA constructs and transfection
As described previously[
18], C-terminal deletion mutants of human CD26 cDNA constructs were generated by using Nco I restriction enzyme sites to delete domain representing the 740-766th AAs in the C terminus, using Nhe I to delete from the 578th AA, using BspE I to delete from the 450th AA, using Stu I to delete from the 359th AA, and using Pst I to delete from the 248th AA. These cDNAs were ligated in-frame into pcDL-SRα expression vector[
20]. The green fluorescence protein (GFP)-expressing vector pEB6-CAG-GFP was a kind gift from Dr. Yoshihiro Miwa (Tsukuba University, Tsukuba, Japan)[
21]. Each CD26 deletion construct in pcDL-SRα was co-transfected with pEB6-CAG-GFP into COS-7 cells using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA). After 24 hours of transfection, cells were harvested, followed by staining with Alexa Fluor 647-labeled 4G8, 1F7, 5F8, YS110, clone 18 or clone 19, and then analyzed by flow cytometry.
Preparation of immunogen
Soluble CD26 (sCD26) was produced according to the method described previously[
22]. Briefly, the expression vector RcSRα-26d3-9, which contains a deletion of the coding sequence for amino acids 3–9 of CD26, was transfected into a dihydrofolate reductase deficit Chinese hamster ovary (CHO) cell line, DXB-11 by electroporation, together with pMT-2 providing the dihydrofolate reductase gene. The transfected CHO cells were cultured in serum-free CHO-S-SFM II medium (Invitrogen) supplemented with 1 μM methotrexate (Nacalai Tesque, Kyoto, Japan). The culture supernatant was collected and subjected to affinity chromatography on ADA-Sepharose according to the method described previously[
23]. Purified sCD26 was denatured in 8 M urea buffer supplemented with 20 mM HEPES and 50 mM dithiothreitol (DTT) by gentle rotation for 8 hours at RT.
Development of hybridomas and monoclonal anti-human CD26 antibodies
Denatured sCD26 was dialyzed in PBS, and 100 μg of protein per 50 μl of PBS was emulsified with 50 μl of adjuvant, TiterMax Gold (TiterMax USA, Norcross, GA). A 6-wk-old female BALB/c mouse was immunized s.c. with 100 μl of the emulsion seven times every two weeks and finally injected i.v. with half volume of the emulsion. Three days after the final immunization, the spleen was removed and 100 × 106 spleen cells were fused with 100 × 106 P3U1 myeloma cells by using polyethylene glycol 4000 (Merck, Darmstadt, Germany) and were cultured in RPMI1640 supplemented with 10% fetal bovine serum (FBS, Japan Bioserum, Fukuyama, Japan), 5% BriClone (NICB, Dublin, Ireland) and HAT (Invitrogen) in 96-well flat-bottom plates (Costar, Corning Incorporated, Corning, NY). Hybridoma supernatants were screened for selective reactivity with human CD26 by using flow cytometry and enzyme-linked immunosorbent assay (ELISA). The supernatants which can detect human CD26 by both flow cytometry and ELISA were finally screened for immunostaining of formalin-fixed paraffin-embedded human tissue sections. The hybridomas were cloned by limiting dilution and culture medium was exchanged for serum-free GIT medium (Wako Pure Chemicals, Osaka, Japan). Monoclonal antibodies were purified from the supernatants using Protein A IgG Purification Kit (Pierce, Rockford, IL) according to the manufacturer’s instructions.
Flow cytometry
A CD26-negative Jurkat T cell line (Jurkat parent) and a stable Jurkat T cell line transfected with human CD26 cDNA (Jurkat-CD26WT) described previously[
24] were used for screening of hybridomas. Cells were washed in PBS containing 1% FBS, 1 mM EDTA and 0.1% sodium azide, and incubated with 100 μl of hybridoma supernatant or 20 μg/ml of purified mouse anti-human CD26 mAb for 25 min at 4°C, and subsequently stained with PE-conjugated goat anti-mouse Ig pAb (BD Biosciences, San Jose, CA) for 25 min at 4°C. Acquisition was performed using FACSCalibur (BD Biosciences) and data were analyzed with FlowJo software (Tree Star, Ashland, OR). For cross-blocking studies of humanized anti-CD26 mAb (YS110), cells were pretreated with unlabeled YS110 or control human IgG (50 μg/ml, respectively) for 25 min at 4°C, and subsequently incubated with 100 μl of hybridoma supernatant or 20 μg/ml of purified mouse anti-human CD26 mAb for 25 min at 4°C, and finally stained with PE-conjugated goat anti-mouse Ig pAb for 25 min at 4°C. For cross-blocking studies of murine anti-CD26 mAbs, cells were pretreated with unlabeled 4G8, 1F7, 5F8, 16D4B, 9C11 or mouse IgG
1 isotype control (50 μg/ml, respectively) for 25 min at 4°C, and subsequently stained with Alexa Fluor 647-labeled clone 18 or clone 19 or PE-conjugated goat anti-mouse Ig pAb for 25 min at 4°C.
ELISA
The 96-well immunoplates (NUNC, Roskilde, Denmark) were coated with native sCD26 or denatured sCD26 described above in carbonate bicarbonate buffer (200 ng/well, respectively) or buffer alone as a negative control at 4°C overnight. Each well of the plate was blocked with 3% bovine serum albumin (BSA, Sigma, St.Louis, MO) in PBS for 1 hour at RT, and then incubated with 3-fold diluted hybridoma supernatants or 5 μg/ml of purified mouse anti-human CD26 mAb or goat anti-human CD26 pAb for 1 hour at RT, and subsequently incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse Ig pAb (BD Biosciences) or HRP-conjugated donkey anti-goat IgG Ab (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 hour at RT. Tetramethylbenzidine (TMB) Peroxidase Substrate (KPL, Gaithersburg, MD) was finally added to each well and the reaction was stopped by 2N H2SO4. The absorbance at 450 nm/570 nm was measured in a Microplate Reader (Bio-Rad, Hercules, CA) and data were analyzed with Microplate Manager 6 software (Bio-Rad).
Tissue specimens and immunohistochemical staining
Formalin-fixed paraffin-embedded tissue specimens of malignant mesothelioma and normal liver, kidney and prostate were used for positive controls in the immunohistochemical examination. The use of human sample from autopsy cases with hepatocellular carcinoma, renal cell carcinoma, prostate adenocarcinoma, colon adenocarcinoma and lung adenocarcinoma was generously permitted by the bereaved families. This study was approved by the Okayama Rosai Hospital ethical review board and the Keio University School of Medicine ethical review board, and the purpose of the study was explained to all patients and their written informed consent was obtained. All studies on human subjects were carried out according to the principles set out in the Declaration of Helsinki. Formalin-fixed paraffin-embedded tissue specimens were cut into 4-6 μm sections and deparaffinized. Antigen retrieval was performed by 1) autoclaving in 10 mM citrate buffer (pH 6.0) for 20 min at 120°C, 2) 0.05% trypsin for 15 min at 37°C, 3) 0.02% proteinase K for 10 min at 37°C, or 4) boiling in 10 mM citrate buffer (pH 6.0) for 10 min at 100°C, and the sections were treated with 0.3% H2O2 in methanol for 10 min at RT to inactivate endogenous peroxidase, then treated with 2.5% horse serum (Vector Laboratories, Burlingame, CA) for 10 min at RT to block non-specific binding of the secondary horse antibody. The sections were treated with 100 μl of hybridoma supernatants or purified mouse anti-human CD26 mAb or goat anti-human CD26 pAb for 2 hours at RT, and subsequently treated with HRP-conjugated horse anti-mouse Ig pAb or HRP-conjugated horse anti-goat IgG pAb (Vector Laboratories) for 30 min at RT. The reaction was visualized with 3, 3’-diaminobenzidine (DAB) (Dojindo Laboratories, Kumamoto, Japan), and the tissue sections were counterstained for nucleus with hematoxylin. To confirm the binding specificity of Abs to human CD26, the anti-human CD26 Ab (100 μg/ml) was gently rotated with 500 μg/ml of sCD26 at 4°C overnight, and after centrifugation, the supernatant was used instead of the primary anti-human CD26 Ab. Expression pattern of CD26 was evaluated and verified independently by two pathologists. The optical microscope images were taken using Axio Scope.A1 microscope (Carl Zeiss, Oberkochen, Germany).
Discussion
Although anti-human CD26 mAbs which we have developed previously or commercially available mAbs cannot clearly detect denatured CD26 in formalin-fixed paraffin-embedded tissues, the anti-human CD26 pAb purchased from R&D Systems is able to stain CD26 with reliable clarity and intensity. However, it is of concern that the staining pattern and intensity may differ among different lots of the anti-CD26 pAb. Since treatment with targeted therapeutic agents depends on the ability to reliably detect the appropriate targets on clinical samples, uniformity of the diagnostic reagents is critical, suggesting that pAbs that are used as research reagents are not appropriate for diagnostic uses in the clinical setting. In the present study, we describe the successful development of novel anti-human CD26 mAbs by immunizing mice with CD26 protein denatured in urea buffer that can potentially be used as diagnostic reagents clinically.
In an attempt to improve diagnostic accuracy, markers used for immunohistochemistry have been studied, such as galectin-3, HBME-1 and CK-19 for diagnosis of benign and malignant thyroid lesions[
25,
26], and FAP-α and Calponin for diagnosing whether ductal carcinoma in situ has microinvasion[
27]. CD26 is highly expressed on the surface of malignant mesothelioma cells especially tumors of the epitheloid and biphasic types, but not on benign mesothelial tissues[
16,
17]. It has been recently reported that the expression level of CD26 in prostate cancer tissues is higher than that of normal prostatic tissues and increased with prostate cancer stage advancement, and CD26 expression is correlated with prostate specific antigen, suggesting that CD26 may be a good marker for prostate cancer diagnosis[
28]. Furthermore, the overall survival of patients with CD26-positive GISTs is worse than that of patients with CD26-negative GISTs, suggesting that CD26 appears to be a reliable biomarker of malignant GIST of the stomach[
29]. These observations strongly suggest that immunohistochemical staining of CD26 in formalin-fixed tumor tissues is important for diagnosis and prognosis of various tumors.
Since several anti-human CD26 mAbs such as Ta1, 1F7, 5F8 and 14D10 that were already developed in our laboratory by immunizing mice with a CD26 positive human T cell line (EL156) or a PHA-stimulated
Aotus trivirgatus T cell line or a murine pre-B human CD26 transfectant (300–19) cannot clearly detect CD26 in formalin-fixed tissues[
1,
8,
18], it was our hypothesis that utilizing human CD26 protein but not human CD26 positive cells as an immunogen would be important for the development of mAbs capable of recognizing the denatured CD26 molecule. Similar to CD26, only pAbs could react to the denatured HLA class I molecules in formalin-fixed paraffin-embedded tissues. Torigoe et al. recently succeeded in developing a novel anti-pan HLA class I mAb suitable for immunohistochemical staining of fixed tissues by immunizing a recombinant HLA-A protein denatured in urea buffer[
30]. The exact role played by urea treatment of the CD26 protein in expanding the repertoire of the obtained anti-CD26 mAs is not yet clear, since we have not examined for potential differences in the characteristics of mAbs obtained after immunizing mice with urea-treated sCD26 protein or non-treated native sCD26 protein in this study. However, as shown in Figures
2 and
3, tissue specimens stained with anti-CD26 mAb purchased from MBL exhibited only a partially positive reaction with weak staining intensity, while this mAb showed higher absorbance to the urea treated sCD26 protein than the absorbance obtained from the novel anti-CD26 mAbs capable of staining CD26 with strong intensity in fixed tissues. These data strongly suggest that the structure of CD26 denatured by the method of antigen retrieval after formalin-fixation is quite different from that of CD26 denatured by urea buffer, and also suggest that anti-CD26 mAbs suitable for immunohistochemistry may be obtained more efficiently by immunizing mice with CD26 protein denatured by methods other than urea treatment, such as treatment with guanidine hydrochloride or sodium dodecyl sulfate (SDS), or with proteases such as trypsin or proteinase K, or by boiling. Further studies are needed to clarify the issue involving pretreatment of the immunogens and the characteristics of mAbs obtained after immunization.
In the present study, we have succeeded in developing novel anti-CD26 mAbs with a wide range of epitopes (Figures
5,
6 and Additional file
1: Figure S3). Since most of these novel mAbs completely inhibited the binding of the anti-CD26 mAbs (4G8, 1F7, 5F8, 16D4B or 9C11) developed previously by our group, the epitopes defined by these novel mAbs are expected to be similar to those recognized by the earlier mAbs. However, these novel anti-CD26 mAbs are capable of detecting denatured CD26 in fixed tissues with strong intensity, unlike the previously developed mAbs. Similarly, while clone 18 and YS110 recognize the similar epitope on CD26 (Figure
4, Additional file
1: Figure S2 and Additional file
1: Figure S3), only clone 18 can stain CD26 clearly in fixed tissues with strong intensity, suggesting that slight differences in the recognized epitopes can determine whether mAb binding to its denatured antigen can occur.
Cross-blocking studies showed that, in contrast to clone 5 or 18, the binding of clone 1, 11, 16 or 19 to CD26 was hardly affected by the humanized anti-CD26 mAb YS110 (Figure
4, Additional file
1: Figure S2 and Additional file
1: Figure S3), suggesting that these 4 novel anti-CD26 mAbs are suitable for analyzing CD26 expression in clinical samples following YS110 therapy. Potential uses of these novel mAbs in the clinical setting may be to detect CD26 expression in formalin-fixed tissues, or on circulating cells in blood samples, or sCD26 levels in bodily fluids through such methods as immunohistochemistry, flow cytometry, or ELISA. Furthermore, these novel mAbs are potentially useful for analyzing CD26 expression in fixed tissues or on the surface of lymphocytes or tumors during or following the administration of humanized anti-CD26 mAb in animal disease models that involve inoculated human lymphocytes or tumors[
12,
16], and are expected to contribute to future CD26-related research effort.
Since we intend to utilize these novel anti-CD26 mAbs as companion diagnostic agents in the clinical setting, our current effort is focused on improving immunohistochemical staining methods by examining such issues as the condition of antigen retrieval or blocking, or the optimal concentration of the primary antibody (anti-CD26 mAb) that can maximize staining intensity while lowering background staining. Furthermore, we also identified the amino acid sequence of the variable region in both the heavy chain and light chain of clone 19 (data not shown), and will aim to refine the ability of this mAb to bind to CD26 through genetic engineering techniques.
In conclusion, we have succeeded in developing novel anti-human CD26 mAbs suitable for immunohistochemical staining of CD26 in formalin-fixed tissue sections with reliable clarity and intensity. Furthermore, since some of these mAbs exhibit no cross-reactivity with the therapeutic humanized anti-CD26 mAb, they are potentially useful as companion diagnostic agents in the clinical setting while advancing future CD26-related research.
Competing interests
The authors declare no competing financial interests associated with this manuscript.
Authors’ contributions
RH and CM designed and coordinated the study. RH, TY, SM and EK conducted the experiments. CM, TY and KO supervised part of the experiments. All authors contributed to the interpretations and conclusions presented. RH and CM wrote the manuscript, and NHD, SI and HY participated in editing it. All authors read and approved the final manuscript.