Background
DPP-IV (CD26) is a cell surface 110 kDa glycoprotein expressed on epithelial cells and leukocyte subsets possessing dipeptidyl peptidase activity. [
1]. The DPP-IV enzyme is known to cleave the N-terminal dipeptide from the incretin hormones glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP). This cleavage inactivates the hormones thereby neutralizing their prandial insulinotropic effect [
2,
3]. Targeting the dipeptidyl peptidase activity with low molecular weight enzyme inhibitors restores incretin activity and has led to the successful development of a DPP-IV inhibitor, sitagliptin, as an effective therapy for Type 2 diabetes [
3]. A concern regarding the potential for DPP-IV inhibitors to affect immune function and increase infection rates has been raised [
4,
5], although a recently published analysis of safety using pooled source data showed no significant difference in the incidence of overall or specific types of infection [
6].
The role of DPP-IV enzymatic activity in immune function has not been extensively studied, however there are a few reports suggesting that DPP-IV can modulate immune responses [
7,
8]. Cell culture studies have implicated DPP-IV as a co-receptor in T cell activation [
1]. In addition, DPP-IV may affect leukocyte trafficking via cleavage of certain chemokines such as SDF-1 [
9]. DPP-IV null animals were shown have reduced humoral immune responses to pokeweed mitogen [
10]. In an Ova asthma model, rats expressing a truncated inactive form of DPP-IV due to a genetic polymorphism were shown to have reduced T cell recruitment to the lungs and decreased Ova-specific IgE titers [
11]. However, studies with DPP-IV deficient animals do not directly address the role of the dipeptidyl peptidase activity as this cell surface protein may possess other non-enzymatic functions [
12‐
14]. In addition, some reports that attributed immunomodulatory effects to DPP-IV enzymatic activity may have been confounded by use of non-selective inhibitors. Indeed, we have previously shown that blockade of T cell activation
in vitro correlates with inhibitor activity directed against DPP8/9 but not against DPP-IV [
15]. Moreover, inhibitors that were previously reported to modulate T cell responses were found to be potent inhibitors of DPP8/9 activity [
16‐
21].
To extend these observations to an in vivo setting in order to better characterize any potential role of DPP-IV in immune function, we investigated the T cell-dependent responses in mice using genetic ablation or pharmacological blockade of DPP-IV. T cell-dependent antibody responses provides a useful model for addressing immune competence as it is dependent on many factors such as antigen processing and presentation, CD4 T cell help, germinal center reactions, B cell activation and differentiation, affinity maturation, and memory cell formation. We report here that genetic ablation or specific inhibition of DPP-IV did not impair T cell-dependent antibody responses. In addition, we find that genetic ablation or specific inhibition of DPP-IV did not compromise cytotoxic T cell function in vivo.
Methods
Mice
Female 8 week old C57Bl/6J and DPP-IV
-/- [
22] mice were obtained from Taconic Laboratory, (Taconic Laboratories, Tarrytown, NY, USA). The DPP-IV
-/- mice were originally obtained from Dr. D Marguet and backcrossed on C57BL/6 to homogeneity [
23]. SNP testing carried out revealed 98.4% B6J background. The knock out animals were generated by mating male and female homozygous null animals. The control animals were age-matched and obtained from the same facility as the null animals. Animal were housed in a specific pathogen-free rodent facility. All animal protocols were approved by the Merck Institutional Animal Care and Use Committee.
Antibodies and reagents
To quantify mouse immunoglobulins by ELISA, the following secondary antibodies were used as per manufactures instructions: Rat anti-mouse lambda-Biotin, anti-mouse kappa-Biotin, Rat anti-mouse IgG1-Biotin (BD Biosciences, San Jose, CA, USA), Goat anti-mouse IgG2a-Biotin, Rat anti-mouse IgG2b-Biotin, Goat anti-mouse IgG3-Biotin, and Rat anti-mouse IgM-Biotin (Southern Biotechnology Associate Inc., Birmingham, AL, USA). (4-hydroxy-3-nitrophenyl) acetyl-chicken γ-globulin (NP-CGG, NP-BSA and Ovalbumin) were obtained from Biosearch Technologies, Novato, CA, USA.
MOG p35–55 was obtained from Sigma-Aldrich, St. Louis, MO, USA. Heat-killed
Mycobacterium tuberculosis was obtained from BD Diagnostics, Franklin Lakes, New Jersey, USA. Pertussis toxin was obtained from List Biological Laboratories, Campbell, CA, USA. The highly selective DPP-IV inhibitor, des-fluro sitagliptin, was synthesized as previously described [
24]. To deliver an effective dose of ~400 mg/kg daily, mice were fed a diet consisting of 6.7 g of this compound per 1 kg Tekland chow (Research diets, New Jersey, USA). The enzyme activity of DPP-IV in the blood was assayed as described earlier [
15].
T cell-dependent antibody responses
Mice were immunized i.p. with 100 μg of NP-CGG in alum for the primary immunization, and 100 μg of NP-CGG in PBS i.p. for the secondary immunization [
25]. Mice were bled via the retro orbital sinus at indicated times and the levels of anti-NP and CGG antibodies and their isotypes were determined by ELISA as described previously [
26,
27]. Briefly, mouse serum antibodies were immobilized onto 96-well plates coated with NP-BSA or CGG and detected with biotin-conjugated anti-mouse immunoglobulins. The assays were developed using streptavidin-europium and plates were read on Victor 2 1420 multilabel counter (Perkin Elmer, Waltham, MA, USA). Relative affinities of serum antibodies were evaluated by using altered ligand density ELISA as described earlier [
27].
Flow cytometry analysis
Cell suspensions prepared from spleens excised from mice on day 9 after immunizations were depleted of erythrocytes by ammonium chloride For four-color cell surface staining, 2 × 106 cells resuspended in PBS containing 2% BSA were incubated with pre-titered dilutions of GL7-FITC, biotinylated anti-IgD, anti-B220-PET-Texas Red, and 2.4 G2-PE for 30 min at 4°C. SA Red 670 was used as a second-step reagent. Cells were analyzed using a FACS Calibur cytometer (Becton-Dickinson, Mountain View, California, USA), and the data were analyzed by FlowJo software (FlowJo, Ashland, Oregon, USA).
Immunohistochemistry
Spleen isolation, flash freezing, sectioning, and immunohistochemistry were all conducted essentially as described previously [
28]. GC numbers were determined by counting the number of PNA
+ structures at 10× magnification.
In Vivo Cytotoxicity Assay
C57BL/6 female and male mice, age 8 ~10 weeks, were used for
in vivo cytotoxicity assays. Female mice were immunized i.p. with syngeneic C57BL/6 male splenocytes (1 × 10
7 in 100 μl PBS) at day 0 and boosted at day 10 to generate anti-H-Y cytotoxic T cells. The
in vivo CTL assay was performed as described earlier with some modifications [
29] Target cells were prepared and adoptively transferred to recipient female mice at day 18 post priming and specific killing of target H-Y
+ cells was analyzed at day 19. Briefly, male and female C57BL/6 naïve splenocytes were isolated and depleted of erythrocyte with RBC lysis buffer (Sigma-Aldrich, St. Louis, MO, USA). The splenocytes were washed 1× with 10% FBS RPMI 1640 and 2× with PBS. The cell densities were adjusted to 2×10
7 cells/ml. Male splenocytes were labeled with a high concentration (5 μM) of 5,6-carboxy-fluorescein succinimdyl ester (CFSE, Molecular Probes, Eugene, OR, USA), and the female splenocyte were labeled with low CFSE (0.5 μM). CFSE labeling was carried out by incubating 2 × 10
7 cells/ml in PBS with CFSE at 37°C in a 5% CO2 incubator for 10 min and quenched by adding 10% FBS RPMI to a final volume of 50 ml. Cells were washed 1× with 10%FBS RPMI and 2× with PBS to remove free CFSE. Male and female splenocytes were mixed 1:1 in PBS for the adoptive transfer. 10
6 cells in 200 μl PBS were injected into recipient female mice via tail vein injection. Similar amounts of cell were injected into the control unimmunized female mice. Spleens from the recipient mice were harvested after 16 hr, and splenocytes were analyzed for specific killing of H-Y
+ target cells by FACS. At least 5,000 CFSE positive cells were analyzed. The recovery and percent killing of the CFSE-labeled targets were calculated as follows:
Data are expressed, as mean ± S.D.
Measurement of antigen-specific T cell recall responses
Anti-MOG responses were induced in 8 to 12-wk-old mice by immunization with MOG p35–55 in Complete Freund's Adjuvant (CFA). A total of 200 μg of MOG p35–55 peptide and 800 mg of killed
Mycobacterium was emulsified in CFA and injected S.C. by means of four injections over the flanks. In addition, 200 ng of pertussis toxin dissolved in 200 μl of PBS was injected i.p. at the day of immunization and again the day after [
30]. The draining lymph node and spleen were extracted at day 9 and cell proliferation assessed with
3[H]-thymidine incorporation upon stimulation with different concentrations of MOG peptide (0, 0.5, 5 and 50 μg/ml). Anti-Ova T cells responses were generated by immunizing the mice with 100 μg of ova in alum in 200 μl volume i.p. The draining lymph nodes and spleen were harvested at day 9 and cells were re-stimulated
in vitro with varying concentrations of ova (0, 0.1, 0.5 and 1 mg/ml). Proliferation was measured as described above. IL-2 in the supernatant was measured by an R&D Systems ELISA kit according to manufacturer's instructions
Statistical analysis
Significant differences were determined using one way ANOVA analysis. A p-value of < 0.05 was considered to be statistically significant.
Discussion
Inhibition of DPP-IV is an effective therapy for type II diabetes. Since this protein is also expressed on the surface of T cells, it is conceivable that DPP-IV inhibitors could have immunomodulatory activity. This possibility was tested by assessing T cell-dependent immune responses in pre-clinical models using both DPP-IV knockout mice and mice treated with a highly selective DPP-IV inhibitor. This inhibitor, des-fluro sitagliptin, has no activity against other dipeptidyl peptidases including DPP-2, DPP-8 and DPP-9 (IC50>10 μM). Several distinct types of assays were used to evaluate in vivo T cell-dependent responses after antigen challenge including measurements of serum antibody titers, T cell recall responses, and cytotoxic T cell killing activity. We found that loss of DPP-IV activity had no significant effect in any of these T cell-dependent immune response assays.
DPP-IV knockout animals did not show any difference in the kinetics, amplitude or quality of antigen specific antibody responses to a hapten or the protein CGG. In addition, the germinal center reaction, affinity maturation and isotype switching of antibodies, all obligately dependent on T cell help, were similar in wild type and DPP-IV knockout animals. Moreover, treatment of animals with a DPP-IV inhibitor did not alter any of the parameters described above showing that DPP-IV enzyme activity is not essential for mounting T cell- dependent antibody responses.
Recently it has been reported that DPP-IV knockout mice T cell exhibited enhanced proliferative responses to a self peptide MOGp35–55 [
30]. Despite using similar experimental procedures, our results failed to reproduce these observations in two independent experiments. In addition, treatment of mice with the DPP-IV inhibitor, des-fluro sitagliptin, did not have any significant effect on MOG-specific T cell recall responses. The reason for this discrepancy is uncertain. Of note, the lack of effect of DPP-IV inhibition on MOG-specific immune responses is similar to results seen for other protein antigens such as CGG and Ova. These results strongly suggest that DPP-IV does not play a significant role in T cell recall responses to protein antigens.
CD8 T cells play an important role in immunity against intracellular pathogens. We investigated the effects of DPP-IV blockade on
in vivo cytotoxic T cell functions by challenging the female mice with male cells to generate H-Y specific cytotoxic T cells. The primed animals challenged with male cells demonstrated equivalent cytotoxicity when the DPP-IV expression was genetically ablated or the DPP-IV enzyme activity blocked with des-fluro sitagliptin in comparison to the wild type animals (Figure
6). These results indicate that the CD8 T cell cytotoxic function is intact in animals with DPP-IV blockade.
Our data add to and extend the earlier observation that the DPP-IV enzyme activity is not required for T cell activation or co-stimulating properties
in vitro [
8,
34‐
37]. Busso
et al have observed no effects of CD26 deficiency on
in vitro or
in vivo proliferation of lymph node and spleen cells nor on
in vivo antibody response to mBSA immunization [
38]. Furthermore, although lower IgG1 and IgG2a titers were found in CD26 knockout animals as compared to wild type mice immunized with poke weed mitogen (PWM) in the primary response [
10] these differences were eliminated upon boosting with PWM. Earlier reports assessing the effect of DPP-IV on T cell functions have used inhibitors that were non-selective against other dipeptidyl peptidases [
16‐
18,
21,
39]. The compound des-fluro sitagliptin used in the current studies is structurally similar to sitagliptin, a highly selective DPP-IV inhibitor that is approved for clinical use in type II diabetic patients. We have previously shown that highly selective DPP-IV inhibitors do not impair T cell activation
in vitro, in contrast to less selective inhibitors that cross-inhibit DPP8 and DPP9 [
15]. The current studies extend these observations by showing that highly selective inhibition of DPP-IV does not disrupt T cell responses
in vivo.
While DPP-IV does not appear to affect T cell-dependent responses, this protein could modulate other types of immune responses. For example, in humans, DPP-IV is reported to associate with adenine deaminase (ADA) on the surface of the T cell and could prevent the inhibitory effects of adenosine on T cell proliferation [
40]. The crystal structures of both the sitagliptin/DPP-IV complex and the ADA/DPP-IV complex have been solved. While sitagliptin binds DPP-IV to inhibit its enzymatic activity [
41], this binding does not alter the overall conformation of the DPP-IV molecule, including the aspect of the molecule that binds to ADA [
42]. Therefore, sitagliptin is highly unlikely to alter the ADA-binding properties of DPP-IV. However, since the CD26 molecule does not associate with ADA in mice its impact on T-cell function cannot be addressed in rodents [
43]. DPP-IV has also been reported to associate with CD45 and could thereby modulate signal transduction [
44]. We did not directly address functions that could be altered by disrupting the cell surface association of CD45 with DPP-IV. Of note, inhibition of DPP-IV enzymatic activity would not be expected to affect the association of DPP-IV with other cell surface proteins. Our studies focus on the role of DPP-IV in
in vivo T-cell functions. Several studies including work done by Morimoto
et al has implicated potentially diverse roles for DPP-IV in human T-cell functions. For instance expression of CD26 on T-cells has been associated with co-stimulatory functions [
45]. Studies performed by this group have used isolated T-cells stimulated
in vitro in a variety of ways (including with anti-CD26 antibody, recombinant soluble CD26, and an Fc-caveolin-1 fusion protein). Indeed, the more recent results from this group focus on dissecting
in vitro protein-protein interactions and signaling by the CD26 molecule and its putative ligand, caveolin-1. DPP-IV was reported to have co-stimulatory activity in studies using isolated T-cells or T-cell lines stimulated by cross-linking surface CD26 with antibodies [
46]. However, the physiological relevance of these studies is uncertain, e.g., endogenous cross-linking ligands have not been identified and, clearly, there is much work to be done to further establish the potential physiologic role that CD26 might play in immune processes.
Isolated human Th1 cells expressed three to six fold more CD26 protein than Th2 cells and were more responsive to anti-DPP-IV cross-linking [
47]. Interestingly, DPP-IV enzyme activities were equivalent in both Th1 and Th2 cells. DPP-IV could have different role in distinct T-cell populations. The potential role of DPP-IV in other T-Cell subsets Th17, Treg requires further studies.
It has been previously shown that DPP-IV is capable of
in vitro cleavage of SDF-1α and a number of other chemokines like CXCL6, CXCL9, CXCL10, CXCL11, CCL3L1, CCL5, CCL11 and CCL22
in vitro [
7,
38,
48‐
54]. Recently it was shown that DPP8 could also cleave CXCL10, CXCL11 and CXCL12
in vitro [
55]. However, whether cleavage of these chemokines occurs
in vivo has yet to be demonstrated and thus the physiological relevance is uncertain. DPP-IV
-/- mice were found to be resistant to G-CSF mediated mobilization of hematopoietic stem cell, suggesting a potential role for DPP-IV in regulating SDF-1α in the bone marrow [
38,
48]. This has led to the postulation that DPP-IV inhibitors could be clinical useful for bone marrow transplantation.
Authors' contributions
KV was involved in conception, design, and analysis, interpretation of experiments, drafting and revision of the manuscript. GP was involved in designing, data acquisition and interpretation of humoral immune response experiment, RP for design, data acquisition and interpretation of T-cell recall and in vivo T-cytotoxic assays. YC contributed to design, data acquisition and interpretation for germinal center and immunoassays, KP ran the DPP IV enzyme assays, GE provided pharmacological support for the compound studies. DZ had input in conception, design and revising it critically for intellectual content. All authors read and approved the final manuscript.