Chemokines mediate the recruitment of leukocytes to the sites of inflammation. N-terminal truncation of chemokines by the protease dipeptidyl peptidase IV (DPPIV) potentially restricts their activity during inflammatory processes such as allergic reactions, but direct evidence in vivo is very rare. After demonstrating that N-terminal truncation of the chemokine CCL11/eotaxin by DPPIV results in a loss of CCR3-mediated intracellular calcium mobilization and CCR3 internalization in human eosinophils, we focused on the in vivo role of CCL11 and provide direct evidence for specific kinetic and rate-determining effects by DPPIV-like enzymatic activity on CCL11-mediated responses of eosinophils. Namely, it is demonstrated that i.v. administration of CCL11 in wild-type F344 rats leads to mobilization of eosinophils into the blood, peaking at 30 min. This mobilization is significantly increased in DPPIV-deficient F344 rats. Intradermal administration of CCL11 is followed by a dose-dependent recruitment of eosinophils into the skin and is significantly more effective in DPPIV-deficient F344 mutants as well as after pharmacological inhibition of DPPIV. Interestingly, CCL11 application leads to an up-regulation of DPPIV, which is not associated with negative feedback inhibition via DPPIV-cleaved CCL11(3–74). These findings demonstrate regulatory effects of DPPIV for the recruitment of eosinophils. Furthermore, they illustrate that inhibitors of DPPIV have the potential to interfere with chemokine-mediated effects in vivo including but not limited to allergy.
Chemokines represent a family of small cytokines that attract and activate leukocytes, resulting in a pivotal role in various diseases (1, 2). In the case of allergic diseases it can be expected that the local concentration of chemokines such as CCL11/eotaxin and the expression of its corresponding receptor, the CCR3 determine to a large extent their biological effects, for example the recruitment of eosinophils. N-terminal truncation of chemokines by peptidases such as dipeptidyl peptidase IV (DPPIV),4 regulating their biological activity, adds another level of complexity to their local action at inflammatory sites, including but not limited to allergy-like reactions of the skin.
DPPIV is outstanding within the group of serine proteases, because it is abundantly distributed, being enzymatically active in both soluble and membrane-bound forms (3). The principle of the inhibition of the serine protease DPPIV to block the degradation of the insulinotropic peptide-hormone GLP-1 has opened up new avenues for the treatment of diabetes type 2 (4). Because DPPIV is not specific for insulinotropic hormones such as GLP-1 and also exerts a modulating effect on a broad range of substrates significantly involved in the regulation of body homeostasis, unwanted effects cannot be excluded (5, 6). Substrates of DPPIV such as neuropeptide Y and substance P are involved in cardiovascular regulation, feeding regulation, and mediation of pain (7, 8). Moreover, DPPIV enzymatic activity potentially modulates immune functions in particular via the processing of chemokines (9).
Thus, in the context of allergic diseases, it is of primary importance to evaluate in vivo, whether the modulation of the local activity of the CCR3 ligand CCL11/eotaxin is affected upon systemic inhibition of DPPIV. Continuous blockade of DPPIV might prolong the activity of locally accumulated CCR3 ligands, ultimately leading to an exacerbation of inflammatory processes. Our data indicate that inhibition of DPPIV increases the mobilization and recruitment of eosinophils thereby promoting allergy-like responses mediated by CCL11 in vivo and give warning of immuno-regulatory dysfunctions in patients receiving DPPIV inhibitors.
Materials and Methods
A DPPIV-mutant F344 rat substrain from Charles River Sulzfeld/Germany [F344/Crl(GER)] (DPPIVneg rats), and a wild-type-like substrain derived from Charles River Portage/USA [F344/Crl(Por)] (DPPIVpos rats) were obtained from a breeding colony kept in barrier-reared conditions as previously described at the Central Animal Laboratory at Hannover Medical School, Germany (10). All research and animal care procedures were approved by the Lower Saxony district government (Hannover, Germany).
Human CCL11 was obtained from PeproTech and n-nonanoyl (NNY)-CCL14 was prepared as previously described (11). CCL11(3–74) was prepared as follows: CCL11 (0.5 mg) was dissolved in 1.5 ml of 0.01 M Tris-HCl buffer (pH 7.4). DPPIV (0.1 U/μl) was added and the reaction was conducted at 37°C. After 2 h the reaction was quenched with 50 μl of trifluoroacetic acid (TFA) and CCL(3–74) was isolated using C18 chromatography. The resulting peptide was analyzed using reverse phase HPLC, electrospray mass spectrometry (MS; purity >90%). The reversible, competitive transition-state analog inhibitor of DPPIV Ile-thia has been previously described (4).
Preincubation of CCL11 or NNY-CCL14, each at a concentration of 10−5 M, was performed for 24 h before biological assays as described previously (12). In addition, in several experiments, the DPPIV inhibitor Ile-thia (1.25 × 10−4 M) was added to samples containing either porcine kidney DPPIV or human serum.
Eosinophils were purified from the venous blood of normal non-atopic healthy or atopic volunteers as described previously (13), all having given their informed consent according to the local research ethics committee at Hannover Medical School.
For the preparation of T cells, PBMCs were isolated from blood by Ficoll gradient centrifugation. PBMCs were then washed and finally resuspended at 7 × 105 cells/ml and cultured as previously described (14). Stably transfected murine pre-B 300.19 cells expressing human CCR3 (15) were obtained from B. Moser (Theodor Kocher Institute, University of Berne, Switzerland).
Normal human epidermal keratinocytes were obtained from Promocell and cultured in the appropriate keratinocyte growth medium with supplement mix (Promocell).
Analysis of intracellular Ca2+ fluxes in stably transfected murine pre-B 300.19 cells expressing human CCR3 was performed on a fluorescence image plate reader (FLIPR) system (Molecular Devices) as previously described (16). CCR3 internalization in eosinophils and flow cytometric measurements were performed as described elsewhere (13, 17). Quantification of CD26 gene expression on keratinocytes using real-time PCR was performed as described previously (18).
Mass spectrometry analysis of CCL11 processing by human T cells
To analyze the processing of the naturally occurring chemokine CCL11, 10 μg of CCL11 were incubated with 1–5 × 106 cultivated human T cells (1 × 106 cells/100 μl) in Dulbecco‘s PBS (pH 7.5) at 37°C and 5% CO2. After 1 and 24 h, samples of 20 or 50 μl were taken and the degradation was stopped with 2 or 5 μl of 2% TFA and frozen to −20°C immediately. Cleavage products of CCL11 were determined offline with a MALDI-MS (Voyager DE-Pro) in linear positive ion mode, as described recently (11).
Intravenous application of CCL11 and measurement of blood eosinophils
In the first set of in vivo experiments, 10 nmol/kg body weight (BW) hCCL11 in 200 μl of saline was injected into the lateral tail vein of DPPIVneg rats (n = 13) vs DPPIVpos rats (n = 12). Blood was taken from the retroorbital sinus of isoflurane anesthetized rats at 1 min before and at 10, 30, and 120 min after injection of CCL11. For the quantification of eosinophils, 200 μl of blood/timepoint were lysed, leukocytes counted, and unfixed cytospins stained by a combined May-Grünwald/Giemsa staining. As previously described, at least 1000 cells were differentiated on each slide by two independent observers (19).
Intradermal application of CCL11 and quantification of eosinophil infiltration
CCL11 was administered intradermally at a dose of 0, 10 (0.084 μg), 100 (0.84 μg), and 1000 (8.4 μg) pmol in 100 μl of saline in DPPIVpos (n = 6/dose) and DPPIVneg (n = 5/dose) rats with two injection sites of the same dose, one on either side of the shaved back-side, respectively. Vehicle (saline) and non-injected skin served as controls. Within the same experimental series, a combined solution of 1000 pmol of CCL11 and the DPPIV inhibitor Ile-thia at a dose of 30 μg (MW: 260.4 g/mol) was injected intradermally using DPPIVpos rats (n = 5). Punch biopsies of 8-mm diameter were always taken under isoflurane anesthesia 4 h after injection of CCL11. Biopsies were either fixed in formalin, embedded in paraffin, and 4-μm sections were prepared or they were snap-frozen and further processed for immunochemistry. Paraffin sections were stained with Giemsa solution, and the eosinophil infiltrates were evaluated. In each section, eosinophils located in the dermis were counted in 50 randomly selected grids, and the number of positive cells (i.e., eosinophils) per cm2 was calculated as previously described (20).
Immunohistochemistry by avidin-biotin complex (ABC) and alkaline phosphatase-anti-alkaline phosphatase (APAAP) methods were performed to provide morphological information on expression and localization of DPPIV protein in snap-frozen skin biopsies derived from naive, vehicle (saline), and CCL11 (1000 pmol) experimental conditions. Sections were processed by incubating for 4 h with anti-rat DPPIV/CD26 mAb (clone 5E8, 1:500, Cell Science) followed by incubation with secondary Abs (1:200 in 5% normal goat serum) either for ABC or APAAP stains as previously described (20, 21). Leukocyte subpopulations recruited to the skin were characterized by double APAAP stainings and quantified as previously described (20). Control sections were included, in which primary Ab was omitted. APAAP stainings appeared to be specific while endogenous peroxidase activity interfered with ABC stainings. Sections from different levels of the APAAP-stained sections were evaluated by two independent observers blinded to the treatment as previously described (20).
DPPIV enzymatic assays
Effect of CCL11(3–74) on DPPIV-catalyzed hydrolysis of glycyl-prolyl-7-amido-4-methylcoumarin (GP-AMC) was analyzed using recombinant human DPPIV expressed in Pichia pastoris (22). Activity of DPPIV was monitored by the release of AMC measured at an excitation/emission of 380/460 nm. Assays were done at 30°C using HEPES buffer pH 7.6 (40 mM) using a total volume of 30 μl in a 384-well microplate. Substrate concentration in the assay was 10 μM and CLL11(3–74) was used in final concentrations of 10−8 to 10−5 M.
For the in vivo study on eosinophil mobilization into blood, the data obtained at different time points were subjected to analysis by ANOVA for repeated measurements using Statview 5.0 software (SAS Institute) with treatment being the between groups factor and repeated measures of cell numbers being the within subject factor. Afterwards, data were analyzed by one-way ANOVAs split by time, if appropriate. For the in vivo study on eosinophil infiltration into skin, one-way ANOVA was performed with either genotype or dosages being the factors. The Fisher protected least significant difference was used for post hoc comparisons. Significant differences (p < 0.05) between controls and test groups are indicated by asterisks with * < 0.05, ** < 0.01, and *** < 0.001. All data are given as arithmetic mean ± SEM.
Purified DPPIV abrogates CCL11-induced CCR3 activation
The reduction of DPPIV-pretreated CCL11-induced calcium mobilization is completely reversed by coincubation of CCL11 with DPPIV and the DPPIV inhibitor isoleucine thiazolidide (Ile-thia) (Fig. 1,A). In contrast to this, 24-h incubation of NNY-CCL14 with DPPIV at 37°C does not affect its ability to mobilize intracellular calcium per se (data not shown). Incubation of CCL11 for 24 h with DPPIV, which results in the truncated form CCL11(3–74), strongly reduces its capacity to internalize the principal CCL11 receptor CCR3 (Fig. 1,B, left middle panel, black bold line). However, a residual activity remains. Again, inhibition of Ile-thia-mediated CCL11 processing by DPPIV prevented the loss of biological activity (Fig. 1,B, left lower panel, black bold line) whereas the activity of NYY-CCL14 is not affected by DPPIV pretreatment (Fig. 1 B, right middle panel).
Effects of CCL11 are down-regulated by human serum containing DPPIV activity
As DPPIV activity might significantly contribute to processing and subsequent inactivation of circulating chemokines in vivo, the ability of human serum to process CCL11 was studied. The processing and its biological relevance was monitored by CCR3 internalization. Incubation of CCL11 with human serum for 24 h resulted in a remarkable loss of activity, which was fully compensated by coincubation with the DPPIV inhibitor Ile-thia (Fig. 1,C, left lower panel). To reinforce the relevance of DPPIV-like serum activity, which is strongly supported by the effects of the inhibitor, and to exclude that no other enzymes are significantly involved in CCL11 processing, NNY-CCL14 was treated under the same conditions. Incubation of NNY-CCL14 with serum did not affect its ability to internalize CCR3 (Fig. 1 C, right middle panel), demonstrating that DPPIV-independent processing does not play a role in its degradation and especially its biological activity.
Cultured CD26+ T cells cleave the N-terminal dipeptide of CCL11
CD26 on the surface of T cells is another DPPIV source, which might contribute to processing of CCL11 in vivo. To investigate the degradation of CCL11 by cell surface-expressed DPPIV, recombinant CCL11 was incubated with cultivated human T cells, expressing CD26/DPPIV. After 1 and 24 h, the reaction was stopped by addition of 2% TFA and analyzed offline using a MALDI-MS. After 1 h, no processing could be observed (Fig. 1,D, upper panel) but after 24 h (Fig. 1 D, lower panel) a main cleavage product of CCL11 (Mav 8360.84 Da) with a molecular mass of 8207.97 Da corresponding to CCL11(3–74) (Mav 8206.68 Da) was detected. In control experiments, a 0.9% sodium chloride solution was added instead of CCL11 and MALDI-MS analysis was conducted. No molecular masses in the range of CCL11 were detected, demonstrating that endogenous CCL11 is not produced under the conditions used. In the presence of the specific DPPIV inhibitor Ile-thia, the degradation of CCL11 to CCL11(3–74) was completely abrogated (data not shown). These data demonstrate that CCL11 can be cleaved to CCL11(3–74) by cultivated human T cells and that this processing is blocked by the DPPIV-specific inhibitor Ile-thia.
CCL11-induced eosinophil mobilization into blood is augmented in DPPIVneg rats
As the in vitro experiments have shown that both soluble DPPIV present in human serum as well as membrane-bound DPPIV expressed on the surface of T cells are processing CCL11 and thereby impairing its biological activity, we were interested to study whether this might be of relevance in the mobilization of eosinophils in vivo. For this issue, two Fischer 344 (F344) rat substrains, one expressing high (DPPIVpos) and the other negligible DPPIV-like activity (DPPIVneg), were used in different experimental settings (10). First, the mobilization of eosinophils into the blood as an initial key step in allergic inflammation was studied to determine a possible role of DPPIV in the regulation of CCL11 participating in this process of inflammation.
The eosinophil fraction of leukocytes in the peripheral blood was determined directly before the i.v. application of CCL11 (10 nmol per kg of body weight) and 10, 30, and 120 min thereafter. The basal eosinophil blood count before injection was 1.0–1.5%. After an initial slight decrease at 10 min after CCL11 injection, the eosinophil numbers increased to a maximum at 30 min in both groups. A significant increase of the number of eosinophils was observed in the DPPIVneg compared with DPPIVpos rats (Fig. 2 A), indicating that CCL11 was more efficient depending on the expression of DPPIV. After 120 min, eosinophils dropped to the levels of the basal eosinophil count again.
Increased numbers of eosinophils are observed after intradermal injection of CCL11 in DPPIVneg rats
As chemokines are known to play a major role in the recruitment of eosinophils to inflammatory sites, it is even more interesting to determine whether the activity of DPPIV also modulates this step in an in vivo setting. We decided to use a simplified model, injecting purified CCL11 into the skin of DPPIVpos vs DPPIVneg rats, to allow the investigation of this situation independent of other species-specific factors, occurring in almost any kind of in vivo disease model.
The basal intradermal content of eosinophils is virtually identical in DPPIVpos and DPPIVneg rats (Fig. 2,B). Injection of the vehicle (saline) does not have any effect and the number of intradermal eosinophils remains equal. Application of CCL11 leads to a dose-dependent increase of infiltrating eosinophils in both of the substrains (Fig. 5, p < 0.0001). In the DPPIVneg rats there was a linear-shaped, significant increase of eosinophil cell counts dependent on the injection doses 10, 100, and 1000 pmol. Similarly, but on a lower level, in the DPPIVpos rats eosinophils significantly increased up to an injection dose of 100 pmol (p < 0.0001). The increase of eosinophils is more pronounced in DPPIVneg rats, resulting in a significantly differential amount of eosinophils in the dermis of DPPIVneg compared with DPPIVpos rats at 1000 pmol (Fig. 2,B), indicating that the regulation of CCL11-dependent eosinophil recruitment to inflammatory sites is affected in a state of DPPIV deficiency at a certain threshold. Furthermore, to validate cleavage effects on CCL11 by DPPIV, the DPPIV-resistant chemokine analog NNY-CCL14 (11) was injected, revealing that a significant increase of infiltrating eosinophils compared with the same dose of CCL11, both injected in DPPIVpos rats, already occurs at a dose of 100 pmol per site as shown in Fig. 2 B (p = 0.01). This effect might depend on the differences in processing by DPPIV.
The DPPIV inhibitor Ile-thia gives rise to increased intradermal eosinophil recruitment
To confirm the findings in the genetic model, we additionally performed an experiment within a pharmacological setting, as a lifelong deficiency of DPPIV is not directly comparable to a limited pharmacological intervention. Fig. 3 summarizes the results of an experiment on the CCL11-dependent recruitment of eosinophils to skin using DPPIVneg and DPPIVpos rats, the latter treated either with saline or with the DPPIV inhibitor Ile-thia along with an intradermal CCL11 application. Compared with DPPIVpos rats receiving saline plus CCL11 injections (Fig. 3,A; see also eosinophils in a representative skin section in Fig. 3,B), application of the DPPIV inhibitor Ile-thia plus CCL11 resulted in significantly increased numbers of intradermal eosinophils (Fig. 3,A and example in Fig. 3 C), which were similar to those found in DPPIVneg rats.
Intradermal injection of CCL11 induces DPPIV expression in the skin and leukocyte recruitment
As local CCL11 application and consecutive inflammatory processes may interact with the expression level of DPPIV, immunohistochemistry on snap-frozen skin biopsies as well as CCL11 coincubation with cultured keratinocytes were performed to study whether and how DPPIV expression is induced in the skin (Fig. 4). There is a low expression of DPPIV in biopsies of untreated and vehicle (saline)-treated skin (Fig. 4, A, B, D, and F). Fig. 4,A provides an overview of epidermis, dermis, and s.c. tissue at low magnification of untreated rat skin while Fig. 4, B, D, and F, illustrate DPPIV immunoreactivity of vehicle-treated skin at higher magnifications of the boxed areas. Interestingly, intradermal injection of CCL11 up-regulates the DPPIV-like immunoreactivity (Fig. 4, C, E, and G). DPPIV immunoreactivity was primarily located in the epidermis, at hair follicles, and at the fascia between dermis and subcutis (Fig. 4, C, E, and G, see arrows), illustrating that local expression of DPPIV is up-regulated by the CCL11-induced inflammatory process in the area surrounding the intradermal injection site. Relative DPPIV gene expression in cultured keratinocytes is not induced by coincubation with CCL11 alone across time but does increase after stimulation with IFN-γ, TNF-α, and IL1β (Fig. 4,H). Apart from eosinophils, also DPPIV/CD26 positive T cells, monocytes, and NK cells are time-dependently recruited to the skin with T cells being initially present at 4h already followed by a peak of monocytes at 8h and a delayed peak of NK cell numbers at 12h (Fig. 4 I). Together these data provide evidence for an indirect induction of DPPIV on keratinocytes via cytokines released by activated cells and not directly via the action of CCL11/eotaxin.
No feedback inhibition of DPPIV enzymatic activity is given by N-terminal truncated CCL11(3–74)
The induction of local DPPIV expression in the vicinity of CCL11 injection sites might inhibit a further progression of the CCL11-induced inflammatory process along with eosinophil recruitment. Thus, it was of interest to establish whether DPPIV-truncated CCL11(3–74) exerts an inhibitory effect on DPPIV-like enzymatic activity, thereby leading to a feedback inhibition and consecutive aggravation of the inflammatory process. In Fig. 5 it is shown that DPPIV is not inhibited by CCL11(3–74) coincubation. No feedback regulation by CCL11(3–74) could be found even within the micromolar range. Thus, N-terminal truncated CCL11(3–74) does not exhibit feedback inhibition of DPPIV enzymatic activity, which is in contrast to many other substrates of DPPIV.
It is now generally accepted that DPPIV acts as an important regulator of multiple physiological processes. It catalyzes the release of dipeptides from the N-terminus of circulating hormones, neuropeptides, and chemokines. Moreover, DPPIV is engaged in T cell-dependent immune responses and has been associated with cell adhesion and tumor metastasis (19, 23, 24). As DPPIV appears to act at a checkpoint of blood glucose homeostasis via potentiation of GLP-1-mediated stimulation of the entero-insular-axis and concomitant release of insulin, it has emerged as a target for the treatment of type 2 diabetes (4). At present, several DPPIV inhibitors are in the late stage of clinical development and some of them have reached the market for this indication (25). However, based on its ubiquitous expression and pleiotropic functions, systemic and continuous pharmacological blockade of DPPIV might act as a double-edged sword, as not only the “beneficial” release of insulin is increased but also immune regulation might be affected in parallel.
The presented data provide direct evidence that under a status of genetically or pharmacologically induced DPPIV deficiency, a prolonged activity of non-processed CCL11 along with CCR3 activation is evident in vitro as well as in vivo, thereby illustrating a potential scenario of unwanted effects during therapy of diabetes type 2. Because CCR3 represents the major chemokine receptor of eosinophils, which themselves are centrally involved in allergic diseases, dysregulation of these cells might be a consequence and problem associated with pharmacological inhibition of DPPIV (26, 27). In fact, FDA-approval of the DPPIV-inhibitor vildagliptin was halted (see www.novartis.com/downloads _new/investors/2006.11.13%20Galvus%20US%20NDA%20Review.pdf), as there was evidence of skin lesions with blistering in one or more monkey tox studies.
As inactivation of chemokines has mostly only been shown in vitro (28), the relevance of DPPIV processing remains questionable. To address this issue in the case of CCL11, we applied an experimental approach in vivo. Recently, it has been shown that CCL11 is able to mobilize eosinophils into blood (29). Mobilization of eosinophils from the bone marrow is thought to represent an initial key step in inflammatory conditions involving eosinophils, such as allergic asthma or atopic dermatitis. The relevance of CCL11 on the mobilization of eosinophils remains contradictory as one group has shown a significant reduction of circulating eosinophils in mice with a targeted disruption of CCL11 compared with wild-type mice (30), while another group found no difference (31). Nevertheless, it is suggested that CCL11 is involved in this process together with the main player IL-5 (32). In guinea pigs, it has been shown that guinea pig CCL11 induces mobilization of eosinophils with a maximal effect 30 min after i.v. CCL11 injection (29). Later on, it has been demonstrated that human CCL11 is able to cause a dose-dependent increase of circulating eosinophils in the BN rat, which is ∼3-fold higher than the vehicle (33). The increased mobilization in F344 rats expressing DPPIV is similar, peaking 30 min after i.v. infusion. CCL11 injection and the kinetics are comparable to those in the guinea pig. In DPPIVneg F344 rats, the mobilization is significantly enhanced at the timepoint of optimal mobilization, providing a first hint that in vivo activity of CCL11 is modulated by DPPIV. The slow degradation of CCL11 by DPPIV questions the in vivo relevance of this process to some extent. However, the degradation of CCL11 has been shown to be completed in the range of one to two and a half hours (9, 11) which is clearly before the time-point (4 h) when the recruitment of eosinophils was quantified in the present experiments. Moreover, it is conceivable, that the DPPIV activity in vivo is even higher, as soluble activity in serum, cell bound activity from T cells and from endothelial cells contribute to overall DPPIV activity. Interestingly, it has been shown that DPPIV activity is found in skin at a significantly higher amount than in serum and that N-terminally truncated CCL11 (desGP) has been purified from supernantants of cultured skin fibroblasts (33, 34).
As a next step, an influence of DPPIV activity on cutaneous CCL11-dependent eosinophil accumulation was investigated. The cutaneous recruitment of eosinophils in rats is supposed to be modulated by DPPIV, as it has been shown that DPPIV is present in rat epidermis (35). Furthermore, the capacity of human CCL11 to attract eosinophils to rat skin has been shown previously (33). A dose-dependent infiltration could be observed in DPPIVpos and DPPIVneg rats. A trend for enhanced cellular recruitment was observed already at the lower concentrations of CCL11 applied in DPPIVneg rats, which became a significant effect at the highest concentration applied, implying that a certain dose is necessary to induce effective cellular recruitment. A similar profile was derived for application of resistant NNY-CCL14 in DPPIVpos rats, substantiating the relevance of DPPIV in limiting the proinflammatory effects of CCL11. To mimic a specific pharmacological intervention and to compare with the findings based on genetic DPPIV deficiency, inhibitor treatment was combined with CCL11 application, revealing also a significant effect at the high dose applied. These results strongly suggest that in fact DPPIV expression represents an important limiting, negative regulating factor in CCL11-induced effects, as illustrated by but not limited to eosinophil recruitment to the skin. Although the DPPIV inhibitor Ile-thia has been shown to also inhibit the intracellularly located DPPIV-related dipeptidyl peptidases DP8 and DP9, the results obtained from the DPPIVneg rats and the in vitro experiments with CCL11 and serum give strong evidence that the observed effects virtually depend on the activity of DPPIV (36).
The relevance of CD26/DPPIV for cellular homing in settings other than inflammatory responses has just recently been impressively demonstrated by Christopherson III and colleagues in the context of hematopoietic stem cell homing and engraftment (37). In the case of bone marrow transplantation, there is a need for augmented homing of cells in contrast to inflammatory disorders. In this study, it was elaborated that inhibition of CD26 on donor cells might be a relevant strategy to enhance the efficiency of transplantation. In contrast to T cells, whose recruitment to inflammatory sites is decreased by DPPIV blockade, CD26-deficient donor cells have increased homing potential. Thus, there are dramatic differences between pathologic condition and type of cells involved with respect to DPPIV activity. A good example suggesting that DPPIV inhibition aggravates eosinophilic inflammation is displayed by the observation that DPPIV activity in serum is decreased in subjects with Churg-Strauss syndrome (38). Although it is not clear whether CCL11 has a specific non-redundant role in allergic inflammation, regulating the mobilization and recruitment of eosinophils, the presented data suggest that modulating the activity of DPPIV with specific inhibitors might have significant impact on many physiological and pathophysiological processes such as inflammation, as many other chemokines and several other biologically relevant peptides are substrates of DPPIV.
Overall, we show here that the CCL11-mediated effects on two steps of eosinophil recruitment in vivo are more effective in DPPIV-deficient F344 rats as well as after pharmacological inhibition of DPPIV. In addition, we illustrate that CCL11 application into skin leads to an up-regulation of DPPIV, which is not associated with negative feedback inhibition via N-terminally DPPIV-cleaved-CCL11(3–74). Together, this demonstrates that the DPPIV-mediated N-terminal truncation of CCL11 limits the recruitment of eosinophils in vivo, which in general exemplifies a critical role of DPPIV for the control of the chemokine network and specifically points to an important role for DPPIV expression and activity in limiting allergic responses.
Nevertheless, so far, the development of DPPIV inhibitors, especially for the treatment of type 2 diabetes, is very promising and the clinical trials, which have been disclosed until recently neither reported on adverse events (39, 40, 41), nor were indicative for a modulation of the chemokine system. However, the biology and the consequences of sustained DPPIV inhibition may be different in subjects with comorbidity such as immune and inflammatory disorders or atopy, which obviously have not yet been studied in sufficient detail in any clinical trials. Therefore, in the future, these aspects should be diligently observed, when DPPIV inhibitors are administered. On the one hand, a transient, intermittent treatment with DPPIV inhibitors might be favorable to avoid putative adverse events associated with the chemokine system. In contrast, the application of longer-acting GLP-1 analogues targeting one specific receptor is an appealing alternative for the treatment of type 2 diabetes (42).
We thank Susanne Fassbender, Susanne Kuhlmann, and Birgit Eilers for their excellent technical assistance.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the HILF-II program of the Hannover Medical School (to S.v.H. and J.E.) as well as by grants from the Deutsche Forschungsgemeinschaft (1. SFB 587; Project B11 to S.v.H., A.S., and J.S.; 2. FO77/10-1 to W.G.F. and J.E.).
Abbreviations used in this paper: DPPIV, dipeptidyl peptidase IV; MS, mass spectrometry; ABC, avidin-biotin complex; APAAP, alkaline phosphatase-anti-alkaline phosphatase; TFA, trifluoroacetic acid.