Clinica Chimica Acta 413 (2012) 456–462

Contents lists available at SciVerse ScienceDirect

Clinica Chimica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /c l inch im

Method comparison of dipeptidyl peptidase IV activity assays and their application inbiological samples containing reversible inhibitors

Veerle Matheeussen a, Anne-Marie Lambeir a, Wolfgang Jungraithmayr b, Nelson Gomez c,Kathleen Mc Entee c, Pieter Van der Veken d, Simon Scharpé a, Ingrid De Meester a,⁎a Laboratory of Medical Biochemistry, University of Antwerp, Antwerp, Belgiumb Division of Thoracic Surgery, University Hospital of Zurich, Zurich, Switzerlandc Laboratory of Physiology, Faculty of Medicine, ULB, Brussels, Belgiumd Laboratory of Medicinal Chemistry, University of Antwerp, Antwerp, Belgium

Abbreviations: DPP, dipeptidyl peptidase; ADA, adeMe-β-NA, glycyl-prolyl-4-methoxy-β-naphthylamide;amino-4-methylcoumarin; Gly-Pro-pNA, glycyl-prolyl-p⁎ Corresponding author. Tel.: +32 3 265 27 41; fax:

E-mail address: ingrid.demeester@ua.ac.be (I. De Me

0009-8981/$ – see front matter © 2011 Elsevier B.V. Alldoi:10.1016/j.cca.2011.10.031

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 22 September 2011Received in revised form 26 October 2011Accepted 26 October 2011Available online 7 November 2011

Keywords:CD26DPP4DPPIV inhibitorsMethod validationPreclinical studiesVildagliptin

Background: Dipeptidyl peptidase IV (DPPIV, DPP4) is a serine protease that releases N-terminal dipeptides. Itis a validated drug target for type 2 diabetes and DPPIV inhibitors are currently evaluated for other therapeu-tic applications. Various assays are used for DPPIV activity measurements in biological samples. Highly sen-sitive methods are needed to measure also very low activities in inhibited samples.Methods: Here, the three most extensively used substrates to quantify DPPIV activity are compared using in-house methods. A luminescent kit was also included. In addition, one of the in-house fluorometric assays waselaborated for use in biological samples containing reversible DPPIV inhibitors to estimate residual DPPIV ac-tivity which is usually underestimated due to sample dilution.Results: The in-house methods showed a good precision, linearity and specificity. Both fluorometric sub-strates had a 10-fold higher sensitivity compared to the colorimetric assay. The luminescent kit was foundto be the most sensitive.Conclusions: All three in-house methods can be used to measure DPPIV activity in non-inhibited biological

samples. The more sensitive fluorometric assays are recommended when sample volumes are limited orwhen using inhibited samples. The elaborated fluorometric method can be used to estimate the residual invivo DPPIV activity in inhibitor treated subjects.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Dipeptidyl peptidase IV (DPP4/DPPIV, EC 3.4.14.5) is a serine pro-tease belonging to the prolyl oligopeptidase family. As a cell surfaceprotease, DPPIV is ubiquitously expressed in almost all organs and tis-sues with the highest local concentration in the proximal tubule ofthe kidney and on the luminal membrane of epithelial cells of thesmall intestine. DPPIV is also known as T-cell activation antigenCD26 or adenosine deaminase (ADA) binding protein. A solubleform of DPPIV (sDPPIV), lacking the cytoplasmic tail and transmem-brane region, is found in plasma and other body fluids like seminalfluid, saliva and bile [1].

DPPIV was discovered in 1966 and named glycylprolylβ-naphthylamidase because it liberates glycylproline resulting inthe formation of β-naphthylamine, which can be detected via

nosine deaminase; Gly-Pro-4-Gly-Pro-AMC, glycyl-prolyl-7-ara-nitroanilide.+32 3 265 26 85.ester).

rights reserved.

colorimetric analysis after an azo dye coupling [2]. After the discover-y, further research revealed that DPPIV cleaves two N-terminal aminoacids from peptides with an alanine or proline at the second positionof the peptide chain. Several functions have already been attributedto DPPIV since it is able to cleave bioactive peptides such as neuro-peptides and chemokines, thereby modifying their biological activity.The most explored function of DPPIV is its role in glucose homeosta-sis. For the moment, several DPPIV inhibitors are therapeutically usedin the treatment of type 2 diabetes. They have been evaluated in sev-eral animal models of disease and were found to have a favorable an-imal safety profile which forced a breakthrough in the developmentof DPPIV inhibitors [3]. The clinically applied inhibitors are thereforeuseful tools to study other features of DPPIV, for example an involve-ment in immune response, ischemia–reperfusion injury, heart failure,cancer, tissue remodeling or neurodegenerative diseases [4,5].

Over the last decades, the enzymatic activity of dipeptidyl pepti-dase IV has been determined in serum samples and tissue homoge-nates of humans as well as laboratory animals by using eitherchromogenic or fluorogenic substrates. The now frequently usedchromogenic p-nitroanilide substrates were introduced in 1976 be-cause of the better suitability for routine use compared to β-

457V. Matheeussen et al. / Clinica Chimica Acta 413 (2012) 456–462

naphthylamine [6]. The need for a fluorescence-based method wasformulated in 1978 to increase the sensitivity of the assay whichmade it possible to detect DPPIV activity also in cerebrospinal fluid[7]. Even since then, the colorimetric assay has still been the most fre-quently used assay since no calibration is needed and no fluorimeteris required. A very sensitive DPPIV activity assay is needed when theavailable volume of biological samples is limited and when activitieshave to be measured in samples after inhibitor treatment. Moreover,the therapeutically used inhibitors, sitagliptin [8], vildagliptin [9],saxagliptin [10] and linagliptin [11], are all reversible inhibitorswhose potency is underestimated in most ex vivo DPPIV activity as-says due to dilution of the samples. This issue has been brought upearlier but was not yet resolved [5,12,13]. The pharmacokinetic deter-mination of the concentration of inhibitor and active metabolites re-quires special equipment and an optimization of the method foreach inhibitor. Therefore, a pharmacodynamic approach, involvingthe measurement of residual DPPIV activity, is a valuable alternativeduring several (pre-)clinical studies of drug candidates.

In this study we first compared 3 kinetic assays, one colorimetricand two fluorometric ones. All were performed on the same appara-tus in microtiterplate format. To the best of our knowledge, this com-parison has never been done before though each of the substrates wasused in published DPP studies [9,14,15]. Such comparison is alsovaluable since various manufacturers have launched several DPPIVactivity kits on the market using different methods. We includedone commercially available DPPIV activity assay kit, using endpointluminescence detection (DPPIV-Glo Protease Assay from Promega).The provider claims that this luminescent reaction is 100 timesmore sensitive than a fluorometric assay. Next to the comparison ofthe different assays, we illustrated the applicability of one of the ki-netic assays to estimate the percentage residual in vivo DPPIV activityin plasma of mice treated with the reversible, slow-binding DPPIV in-hibitor, vildagliptin [16].

2. Materials and methods

2.1. Fluorometric assay using glycyl-prolyl-4-methoxy-β-naphthylamide

DPPIV enzymatic activity was assayed by using glycyl-prolyl-4-methoxy-β-naphthylamide (Gly-Pro-4-Me-β-NA) as fluorogenic sub-strate [17]. In a 96-well plate 10 μl plasma samples were mixed with100 μl 0.5 mM Gly-Pro-4-Me-β-NA (diluted from a 100 mM stockso-lution in DMSO, stored at −20 °C) in 50 mM Tris buffer pH 8.3 in afinal volume of 110 μl. DPPIV activity was determined kinetically dur-ing 10 min at 37 °C by measuring the velocities of 4-Me-β-NA release(λex=340 nm, λem=430 nm) from the substrate. Fluorescence in-tensity was related to a 4-Me-β-NA (stock solution: 20 mM inDMSO, stored at −20 °C) standard curve in the same buffer.

2.2. Fluorometric assay using glycyl-prolyl-7-amino-4-methylcoumarin

DPPIV enzymatic activity was assayed by using 0.5 mM glycyl-prolyl-7-amino-4-methylcoumarin (Gly-Pro-AMC) as fluorogenicsubstrate under the same assay conditions as described above. Thevelocity of AMC release from the substrate is measured at the follow-ing wavelengths, λex of 380 nm and λem of 460 nm. Gly-Pro-AMC andAMC were purchased from Bachem (Weil am Rhein, Germany).

2.3. Colorimetric assay using glycyl-prolyl-para-nitroanilide

To measure the DPPIV activity colorimetrically, glycyl-prolyl-para-nitroanilide (Gly-Pro-pNA) was used as chromogenic substrate, as de-scribed earlier [14]. In a 96-well plate 10 μl plasma samples weremixed with 0.5 mM Gly-Pro-pNA in the same Tris buffer as for thefluorometric assays but in a final volume of 200 μl. DPPIV activitywas determined kinetically during 10 min at 37 °C by measuring the

velocities of pNA release (405 nm) from the chromogenic substrate.The molar absorptivity of pNA at 405 nm amounts 10.2 mM−1 cm−1

and the pathlength of a volume of 200 μl per 96-well is 0.5761 cm.

2.4. DPPIV-Glo™ protease assay

The assay (Promega, Leiden, The Netherlands) was performedaccording to the manufacturer's instructions, being at room tempera-ture. To calculate DPPIV activity (U/L) out of the relative light units(RLU), a standard curve of purified DPPIV with known activities (asmeasured by the colorimetric assay also performed at room tempera-ture) was included. Since the exact formulation of the assay bufferwas not disclosed, we measured the pH and determined it to be 8.3.Therefore, samples were diluted in a 50 mM Tris–HCl buffer pH 8.3.

One unit of enzymatic activity is defined as the amount of enzymethat catalyses the release of 1 μmol of 4-Me-β-NA, AMC or pNA fromthe substrate per minute under the assay conditions. For all assaysthe Infinite™ 200 (Tecan Group Ltd., Männedorf, Switzerland) wasused to detect the velocities of either 4-Me-β-NA, AMC or pNA re-lease. The values of the Michaelis constant (KM) were determinedby plotting the initial velocities of product formation against at least6 different substrate concentrations. The data were fitted to theMichaelis–Menten equation by non-linear regression analysis usingGraFit software version 5 (Erithacus Software Limited, Horley, UK).All reagents were purchased from Sigma-Aldrich (Bornem, Belgium)unless stated otherwise. Human DPPIV was purified as described ear-lier [18]. Sitagliptin was extracted from Januvia™ tablets (Merck,Brussels, Belgium) in a three-step procedure. First, sitagliptin's prima-ry amine function was derivatised with a tert-butyloxycarbonyl(Boc-) protecting group using Boc-anhydride. Insoluble tablet matrixcomponents were removed by filtration and the Boc-derivatised sita-gliptin was further purified and isolated as a single compound usingflash chromatography. Sitagliptin was liberated by treatment with tri-fluoroacetic acid (TFA). The TFA-salt of sitagliptin obtained was trans-formed into an HCl salt. This salt was then obtained in a pure formafter recrystallisation. The identity of sitagliptin HCl was confirmedusing NMR-spectroscopy and ESI-mass spectrometry and its puritydetermined as >98% using HPLC and LC–MS. Concentrations were de-termined by weight.

2.5. Method comparison

To compare the methods, linear regression analysis was per-formed using Grafit software, Bland Altman plots were constructedwith MedCalc software version 11.6.1 (MedCalc Software bvba,Mariakerke, Belgium).

2.6. Procedure for estimating the residual DPPIV activity using thefluorometric assay

To estimate the residual DPPIV activity in biological samples of in-hibitor treated subjects, one can create a calibration curve withknown concentrations of the inhibitor diluted in the biological sam-ple one wishes to analyze. The percentage residual DPPIV activitycan be calculated by comparing DPPIV enzymatic activity of treatedsubjects with non-treated subjects (defined as 100% activity). Forthis study the calibration curve was constructed for vildagliptin inmurine plasma since mousemodels are largely used to study other ef-fects of DPPIV inhibitors in vivo. A stock solution of 100 mM vildaglip-tin (custom synthesized by GLSynthesis Inc., Worcester) in PBS wasdiluted in mouse (C57BL/6) plasma in different concentrations, rang-ing from 1 nM to 3 mM (final concentrations in assay: 0 nM,0.1 nM–300 μM). The resulting DPPIV activity was assayed by usingthe fluorometric assay (described above) with slight modifications.Ten μl of inhibitor diluted in plasma was pre-incubated during15 min at 37 °C with 50 μl 50 mM Tris buffer pH 8.3 to allow

Table 1Analytical performance of 2 fluorometric DPPIV assays (using Gly-Pro-4-Me-β-NA orGly-Pro-AMC as substrate) and 1 colorimetric assay (using Gly-Pro-pNA as substrate).Sensitivity is defined by limit of detection (LOD) and limit of quantification (LOQ). Pre-cision is defined as within-run and between-run variations.

In-house fluorometricassay: Gly-Pro-4-Me-β-NA

In-housefluorometric assay:Gly-Pro-AMC

In-housecolorimetric assay:Gly-Pro-pNA

Sensitivity U/L U/L U/LLOD 0.10 0.18 1.56LOQ 0.26 0.27 2.92

Precision % % %Within-run 2.1–3.1 1.7–3.1 2.2–4.3Between-run 1.3–3.3 1.5–3.5 2.0–4.7

458 V. Matheeussen et al. / Clinica Chimica Acta 413 (2012) 456–462

equilibration. Thereafter, 50 μl of the fluorogenic substrate Gly-Pro-4-Me-β-NA was added to a final concentration of 0.5 mM. The release of4-Me-β-NA (λex=340 nm, λem=430 nm) from the substrate wasdetermined kinetically during 10 min at 37 °C. The resulting DPPIVactivity in the samples with different inhibitor concentrations wasexpressed as the percentage activity versus the non-inhibited sample.The percentage DPPIV activity was plotted versus the concentration ofinhibitor using a non-linear fit model of GraFit 5 software. The equa-tion of the resulting graph was used to estimate the percentage in vivoDPPIV inhibition in plasma samples of mice treated with vildagliptin.

3. Results

3.1. KM values for the 3 substrates

The KM value for DPPIV of the 2 fluorescent substrates and thechromogenic substrate was determined on the same day using thesame purified DPPIV dilution to exclude between day variations.The KM values for the fluorogenic substrates were comparable,60.1±4.7 μM and 58.8±1.3 μM for Gly-Pro-4-Me-β-NA and -AMCrespectively. For the chromogenic substrate, the KM value was almosttwice as high, 101.2±4.2 μM. For all three substrates a final concen-tration of 500 μM was chosen in the assay to measure the activitiesclose to the Vmax to obtain zero order kinetics.

3.2. Performance of both the fluorometric and colorimetric assays

3.2.1. SensitivityThe limit of detection (LOD), defined as the mean result of the

assay plus 3 SD in a series of 20 blank solution runs, was 0.10 and0.18 U/L for the fluorometric (Gly-Pro-4-Me-β-NA and Gly-Pro-AMCrespectively) and 1.56 U/L for the colorimetric assay. The lower limitof quantification (LOQ), defined as the mean result of the assay plus10 SD in a series of 20 blank solution runs, was 0.26 and 0.27 U/Lfor the fluorometric (Gly-Pro-4-Me-β-NA and Gly-Pro-AMC respec-tively) and 2.92 U/L for the colorimetric assay (Table 1). The sensitiv-ity with Gly-Pro-4-Me-β-NA is much better in comparison with the

Table 2A summary of the characteristics of commercially available DPPIV activity assays.

Assay Company

DPPIV drug discovery kit Enzo life sciencesInnoZyme™ DPPIV activity assay kit MerckDPP (IV) inhibitor screening assay kit AbnovaDPPIV-Glo™ protease assay PromegaInnoZyme™ DPPIV immunocapture activity assay MerckFluorogenic DPP4 assay kit BPS BioscienceDPPIV/CD26 assay kit for biological samples Enzo life sciences

original paper, using this substrate but with detection of the fluores-cence at the end-point instead of kinetically, where 5.0 U/L was deter-mined as the limit of detection [17]. A 10 fold lower sensitivity for thecolorimetric assay was achieved compared to the fluorometricmethods under these assay conditions. The sensitivity was also de-fined for the DPPIV-Glo Protease Assay and an LOD of 0.035 U/L andLOQ of 0.040 U/L was measured. This luminescent method was asexpected more sensitive than the others. However, the 100 foldhigher sensitivity, as described by the manufacturer of the assay,was not confirmed in our experiments. All tested methods are usefulfor measuring DPPIV activities in non-inhibited serum because for ex-ample in humans the DPPIV serum activity ranges from 12.5 to42.0 U/L [19].

3.2.2. PrecisionThe within-run coefficient of variation of three dog plasma sam-

ples with various DPPIV activities, measured several times on thesame day (n=10), ranged from 2.09 to 3.13% for Gly-Pro-4-Me-β-NA, from 1.67 to 3.05% for Gly-Pro-AMC and from 2.19% to 4.29% forthe chromogenic substrate. The between-run variation coefficient inseven samples measured on different days (n=20) was 1.32–3.32%,1.47–3.50% and 1.99–4.74% respectively (Table 1). Both within- andbetween-run variations were lower for one of the in-house assayscompared to a commercially available kit from Abnova (Table 2)using Gly-Pro-AMC as substrate and reporting variations of 3.9 and4.1% respectively. A better between-run variation is achieved for theGly-Pro-4-Me-β-NA assay in comparison with the original paper de-scribing variations up to 6.4% [17].

3.2.3. LinearityThe activity of 7 dilutions (1:1, 1:2, 1:4, 1:8, 1:16, 1:32, and 1:64) in

both BSA (70 g/l) or PBS of 4 dog plasma samples was determined. Inthis activity range from 1 to 90 U/L linear regression analysis showedexcellent linearity for both fluorometric assays (r2=0.998±0.001 forBSA dilutions and r2=0.995±0.006 for PBS dilutions using Gly-Pro-4-Me-β-NA and r2=0.998±0.002 for BSA dilutions andr2=0.997±0.003 for PBS dilutions using Gly-Pro-AMC) and the color-imetric assay (r2=0.999±0.001 for BSA dilutions and r2=0.999±0.001 for PBS dilutions).

DPPIV activity measurements in a series of mixtures preparedfrom 2 plasma samples with low and high activity, yielded DPPIV ac-tivities that did not differ withmore than 5% from the expected valuesfor the three substrates.

The dynamic range was determined by spiking a plasma pool withpurified human DPPIV. Activities ranging from 30 to 400 U/L wereused. All in-house assays showed excellent linearity in the examinedconcentration range (r2=0.999 for both fluorogenic substrates and0.998 for the colorimetric assay).

3.2.4. SpecificityTo attribute the Gly-Pro liberating activity solely to DPPIV and not

the more recently discovered and related enzymes DPP8 or DPP9[20,21] or other unidentified factors, the selective DPPIV inhibitor,

Substrate Use

Gly-Pro-pNA/Gly-Pro-AMC DPPIV inhibitor screeningGly-Pro-AMC DPPIV inhibitor screeningGly-Pro-AMC DPPIV inhibitor screeningGly-Pro-aminoluciferin DPPIV inhibitor screeningGly-Pro-AMC Human biological samplesAMC conjugated substrate Purified DPPIV or cell extractsGly-Pro-pNA/Gly-Pro-AMC Biological samples

459V. Matheeussen et al. / Clinica Chimica Acta 413 (2012) 456–462

sitagliptin, was added in a final concentration of 10 μM. In all threemethods, over 98% of the total activity measured could be inhibitedby sitagliptin, reflecting the excellent specificity of these threemethods when measuring DPPIV activities in plasma. This finding isin accordance with an earlier report where at least 95% of the serumGly-Pro liberating activity was attributed to DPPIV based on ADA-binding properties and structural characteristics [1].

A

B

C

0

DPP

IV a

ctiv

ity f

luor

omet

ric

Gly

-Pro

-4-M

e-be

ta-N

A (

U/L

)

0

20

40

60

80

0

DPP

IV a

ctiv

ity f

luor

omet

ric

Gly

-Pro

-AM

C (

U/L

)

0

20

40

60

80

0

DPP

IV a

ctiv

ity lu

min

esce

ntD

PPIV

-Glo

pro

teas

e as

say

(U/L

)

0

20

40

60

80

100

Rat

io o

f co

lori

met

ric

(pN

A)

and

R

atio

of

colo

rim

etri

c (p

NA

) an

d

Rat

io o

f co

lori

met

ric

(pN

A)

and

20 40 60 80 100

20 40 60 80 100

20 40 60 80 100

Gly-Pro-pNA (U/L)

Gly-Pro-pNA (U/L)

Gly-Pro-pNA (U/L)DPPIV activity colorimetric

DPPIV activity colorimetric

DPPIV activity colorimetric

Fig. 1. Correlation between different methods measuring DPPIV activity in dog plasma sa(A) Fluorometric substrate Gly-Pro-4-Me-β-NA, (B) fluorometric substrate Gly-Pro-AMC anare compared with the widely used colorimetric method using Gly-Pro-pNA as a chromoge

3.3. Method comparison

Hundred dog plasma samples were measured with both fluoro-genic substrates and the chromogenic substrate. Twenty of thesesamples were also measured using the Promega DPPIV-Glo ProteaseAssay. Both the fluorometric assays and the luminescent assay werecompared with the most widely used colorimetric assay. All methods

00.9

1.0

1.1

1.2

1.3

1.4

Mean

1.16

-1.96 SD

1.02

+1.96 SD

1.30

00.9

1.0

1.1

1.2

1.3

Mean

1.11

-1.96 SD

1.01

+1.96 SD

1.21

00.8

0.9

1.0

1.1

1.2

Mean

1.03

-1.96 SD

0.91

+1.96 SD

1.14

fluo

rom

etri

c (4

-Me-

β-N

A)

fluo

rom

etri

c (A

MC

) lu

min

escn

t (D

PPIV

-Glo

)

luminescent (DPPIV-Glo)

fluorometric (AMC)

20 40 60 80 100

20 40 60 80 100

Average of colorimetric (pNA) and

Average of colorimetric (pNA) and

fluorometric (4-Me-β-NA) Average of colorimetric (pNA) and

20 40 60 80 100

mples, analyzed by a linear regression model and evaluation of a Bland-Altman plot.d (C) luminescent detection using the DPPIV-Glo Protease Assay kit. All three methodsnic substrate.

460 V. Matheeussen et al. / Clinica Chimica Acta 413 (2012) 456–462

tested showed a good correlation with the colorimetric assay. Thiswas demonstrated by linear regression analysis and evaluation ofBland–Altman plots [22]. Linear regression analysis of Gly-Pro-4-methoxy-β-NA with Gly-Pro-pNA showed a curve with the followingequation: y=0.85x+0.55 (r2=0.98) (Fig. 1A). Other equations arey=0.89x+0.43 (r2=0.99) and y=0.96x+0.57 (r2=0.99) for Gly-Pro-AMC and DPPIV-Glo Protease Assay respectively (Fig. 1B and C).

3.4. Estimation of the residual in vivo DPPIV activity

First, a calibration curve for different concentrations of the revers-ible DPPIV inhibitor, vildagliptin, in mice plasma was constructed.These activity measurements resulted in a curve with the followingequation: y ¼ Yrange

1þ xIC50

� �s þ background or y ¼ 94:01þ x

0:20ð Þ0:58 þ 6:7.

With x=final concentration of vildagliptin in the assay (μM),y=percentage DPPIV activity compared to a non-inhibited sample(%), s=slope factor and IC50=the half maximal inhibitory concentra-tion (μM). Second, the percentage in vivo inhibition in plasma of micetreated with vildagliptin can be calculated via this calibration curve.For example, a plasma sample which shows in the assay 60% of theDPPIV activity when compared to untreated/uninhibited plasma hasa total concentration of 0.13 μM in the assay where it was 11 fold di-luted. So the concentration of vildagliptin in the sample is 1.4 μM.Such a concentration is estimated to be able to inhibit 70% of theDPPIV plasma concentration and thus results in 30% residual in vivoDPPIV activity (Fig. 2). This residual in vivo activity (30%) is lowerthan the activity measured ex vivo in the assay (60%). It can be esti-mated for every inhibitor treated animal to study the pharmacody-namics of inhibitors.

4. Discussion

The Gly-Pro-based fluorometric and colorimetric assays to deter-mine DPPIV activity all perform well, concerning precision, linearityand specificity. Nonetheless, a higher sensitivity was achieved withthe fluorometric substrates. This finding does not necessarily assignthe fluorometric assays as the primary choice because DPPIV activity

concentr

DPP

IV a

ctiv

ity (

%)

0

20

40

60

80

100

estimated percentage

DPPIV activity

1

3

4

DPPIV activity

percentage in vitro

residual percentage in vivo DPPIV activity

in vivo inhibition of 10-4 10-3 10

concentration vildag

Fig. 2. Calibration curve of DPPIV activity, measured by a fluorometric assay (Gly-Pro-4-Me-ma. This graph can be used to estimate the percentage in vivo inhibition of DPPIV activity i

in non-inhibited samples can easily be measured by all three of thein-house methods. When biological sample volumes are limited orin biological samples with a very low DPPIV activity or with anexpected inhibition of the DPPIV activity due to treatment with an in-hibitor, the fluorometric assays might be preferable.

Commercially available DPPIV activity kits all use the substrateswe tested here, except the Promega DPPIV-Glo Protease Assaywhich uses a luminescent detection method (Table 2). We includedthis kit in our method comparison study because Gly-Pro-aminoluciferin as a substrate is not commercially available on itsown and therefore an in-house method cannot be created. Both thefluorometric assays and this luminescent kit showed a good correla-tion with the colorimetric assay. Most of the commercial kits aredesigned for screening of DPPIV inhibitors, also the luminescent Pro-mega kit. This luminescence method was even more sensitive thanthe fluorometric substrates but we did not reach a 100 fold increasein sensitivity as quoted by the manufacturer. The higher sensitivityof this method compared to the others allows the use of loweramounts of DPPIV in drug screenings or of small biological samplevolumes in case of limited sample availability. As shown here, an op-timized in-house DPPIV activity method largely fulfills all require-ments and is less expensive. Both for screening of DPPIV inhibitorsand activity measurements in biological samples, in-house DPPIV as-says are cheaper even taken into account the price of commercialDPPIV preparations. To express the DPPIV enzyme activity of biologi-cal samples in units per liter using the luminescent kit, a sample withknown DPPIV activity must always be included.

By using sitagliptin, we demonstrated that in normal plasma 98%of the Gly-Pro liberating activity is attributed to DPPIV. In otherbody fluids, tissue homogenates and cell extracts, related dipeptidylpeptidases might contribute to the Gly-Pro liberating activity andtherefore it is recommended to check specificity. The same is truefor plasma samples with extremely high DPPIV activity. The inhibitorP32/98 is added in the DPPIV activity kit for biological samples fromEnzo life sciences. The limited specificity of this inhibitor raises ques-tions to its applicability [23]. Sitagliptin is better suited for this pur-pose. Another kit (InnoZyme DPPIV immunocapture activity assayfrom Merck) uses immunocapture of DPPIV before measuring the

ation vildagliptin (µM) 1

11 fold dilution in assay

2

-2 10-1 101 102

liptin in assay concentration vildagliptin in plasma

β-NA), in the presence of different concentrations of vildagliptin diluted in mouse plas-n plasma samples individuals treated with this reversible DPPIV inhibitor.

461V. Matheeussen et al. / Clinica Chimica Acta 413 (2012) 456–462

DPP activity which also seems an effective strategy to distinguish be-tween DPPIV and other activities. A disadvantage of this approach isthat this kit can only be used for analysis of human samples, sinceno information concerning crossreactivity of the antibody withother species is included. The recommendation to perform the activ-ity assay at room temperature given that DPPIV retains more activitythan DPP8 and DPP9 when the assay temperature is dropped to roomtemperature (according to the manufacturer of one of the kits), is ofcourse only a semi-quantitative measure.

We also present a method to estimate the percentage in vivo inhi-bition by reversible DPPIV inhibitors in plasma samples. We thereforechoose one of the fluorometric substrates, namely Gly-Pro-4-Me-β-NA because of its higher sensitivity compared to the chromogenicsubstrate. Such an estimation method is needed in case of DPPIV in-hibitor treatment because the inhibition by reversible inhibitors willotherwise be underestimated due to the dilution of the samples inthe assay. This method estimates the in vivo inhibition based on theconstruction of a calibration curve with different concentrations of in-hibitor in plasma. It can be used for dose-finding of DPPIV inhibitorsin different laboratory animals, especially when long-term inhibitionof DPPIV activity is aimed. Calibration curves as demonstrated inFig. 2 should be constructed for every inhibitor and diluted in anytype of biological sample one wishes to analyze. Although the thera-peutically used inhibitors are all competitive reversible inhibitorsand are shown to have a similar efficacy (i.e. maximal effect) for inhi-bition of DPPIV in vitro, they differ in their potency towards DPPIV (i.e.concentration of compound needed to obtain a defined pharmacody-namic outcome), reviewed in [5]. This should be considered wheninterpreting results since a similar pharmacodynamic outcome (i.e.effect on in vivo DPPIV activity) for 2 different inhibitors, does notimply an identical concentration of both inhibitors in the samples.The presented method has been used by our research group fordose-finding in mice, rats and pigs, in both plasma and serum, for vil-dagliptin as well as sitagliptin and for different administration routes[24]. The suitability of the method for measurements in both serumand plasma was expected since a good correlation between theDPPIV activities in serum and citrate plasma has been demonstrated[19]. Recent studies with DPPIV inhibitors in animal models oftendo not mention the inhibition of DPPIV activity [25–35] or whenDPPIV activity is measured, they do not describe the method[36–40] or describe the method so concisely that it is a challenge torepeat it [41]. Therefore our method might serve as a standard for fu-ture research.

In conclusion, all three, the colorimetric and both fluorometric, in-house methods can be used to measure DPPIV activity in non-inhibited biological samples. The more sensitive fluorometric assaysare recommended when sample volumes are limited or when usinginhibited samples. In addition, the elaborated fluorometric methodcan be used to estimate the residual in vivo DPPIV activity in inhibitortreated subjects.

Acknowledgments

This work was supported by the Fund for Scientific Research Flan-ders (Belgium, FWO-Vlaanderen, grant no. G016209) and by BOF GOAUA (Research Council, Special Fund for Research, University of Ant-werp). Veerle Matheeussen is a research assistant of FWO-Vlaanderen.

References

[1] Durinx C, Lambeir AM, Bosmans E, et al. Molecular characterization of dipeptidylpeptidase activity in serum: soluble CD26/dipeptidyl peptidase IV is responsiblefor the release of X-Pro dipeptides. Eur J Biochem 2000;267:5608–13.

[2] Hopsu-Havu VK, Glenner GG. A new dipeptide naphthylamidase hydrolyzingglycyl-prolyl-beta-naphthylamide. Histochemie 1966;7:197–201.

[3] De Meester I, Scharpé S, Lambeir AM. Dipeptidyl peptidases and related pro-teins: multifaceted markers and therapeutic targets. Clin Chem Lab Med2009;47:245–7.

[4] Lambeir AM, Scharpe S, De Meester I. DPP4 inhibitors for diabetes—what next?Biochem Pharmacol 2008;76:1637–43.

[5] Deacon CF. Dipeptidyl peptidase-4 inhibitors in the treatment of type 2 diabetes: acomparative review. Diabetes Obes Metab 2011;13:7–18.

[6] Nagatsu T, Hino M, Fuyamada H, Hayakawa T, Sakakibara S. New chromogenic sub-strates for X-prolyl dipeptidyl-aminopeptidase. Anal Biochem 1976;74:466–76.

[7] Kato T, Nagatsu T, Kimura T, Sakakibara S. Fluorescence assay of x-prolyldipeptidyl-aminopeptidase activity with a new fluorogenic substrate. BiochemMed 1978;19:351–9.

[8] Kim D, Wang L, Beconi M, et al. (2R)-4-oxo-4-[3-(trifluoromethyl)-5,6-dihydro[1,2,4]triazolo[4,3-a]pyrazin −7(8H)-yl]-1-(2,4,5-trifluorophenyl)butan-2-amine: a potent, orally active dipeptidyl peptidase IV inhibitor for the treatmentof type 2 diabetes. J Med Chem 2005;48:141–51.

[9] Villhauer EB, Brinkman JA, Naderi GB, et al. 1-[[(3-hydroxy-1-adamantyl)amino]acetyl]-2-cyano-(S)-pyrrolidine: a potent, selective, and orally bioavailable dipep-tidyl peptidase IV inhibitor with antihyperglycemic properties. J Med Chem2003;46:2774–89.

[10] Augeri DJ, Robl JA, Betebenner DA, et al. Discovery and preclinical profile of Sax-agliptin (BMS-477118): a highly potent, long-acting, orally active dipeptidyl pep-tidase IV inhibitor for the treatment of type 2 diabetes. J Med Chem 2005;48:5025–37.

[11] Eckhardt M, Langkopf E, Mark M, et al. 8-(3-(R)-aminopiperidin-1-yl)-7-but-2-ynyl-3-methyl-1-(4-methyl-quinazolin -2-ylmethyl)-3,7-dihydropurine-2,6-dione (BI 1356), a highly potent, selective, long-acting, and orally bioavailableDPP-4 inhibitor for the treatment of type 2 diabetes. J Med Chem 2007;50:6450–3.

[12] Edmondson SD, Mastracchio A, Duffy JL, et al. Discovery of potent and selectiveorally bioavailable beta-substituted phenylalanine derived dipeptidyl peptidaseIV inhibitors. Bioorg Med Chem Lett 2005;15:3048–52.

[13] Pospisilik JA, Stafford SG, Demuth HU, et al. Long-term treatment with the dipep-tidyl peptidase IV inhibitor P32/98 causes sustained improvements in glucose tol-erance, insulin sensitivity, hyperinsulinemia, and beta-cell glucoseresponsiveness in VDF (fa/fa) Zucker rats. Diabetes 2002;51:943–50.

[14] Dubois V, Lambeir AM, Van der Veken P, et al. Purification and characterization ofdipeptidyl peptidase IV-like enzymes from bovine testes. Front Biosci 2008;13:3558–68.

[15] Matheeussen V, Baerts L, De Meyer G, et al. Expression and spatial heterogeneityof dipeptidyl peptidases in endothelial cells of conduct vessels and capillaries. BiolChem 2011;392:189–98.

[16] Brandt I, Joossens J, Chen X, et al. Inhibition of dipeptidyl-peptidase IV catalyzedpeptide truncation by Vildagliptin ((2S)-{[(3-hydroxyadamantan-1-yl)amino]acetyl}-pyrrolidine-2-carbonitrile). Biochem Pharmacol 2005;70:134–43.

[17] Scharpé S, De Meester I, Vanhoof G, et al. Assay of dipeptidyl peptidase IV inserum by fluorometry of 4-methoxy-2-naphthylamine. Clin Chem 1988;34:2299–301.

[18] De Meester I, Vanhoof G, Lambeir AM, Scharpé S. Use of immobilized adenosinedeaminase (EC 3.5.4.4) for the rapid purification of native human CD26/dipepti-dyl peptidase IV (EC 3.4.14.5). J Immunol Methods 1996;189:99–105.

[19] Durinx C, Neels H, Van der Auwera JC, Naelaerts K, Scharpe S, De Meester I. Refer-ence values for plasma dipeptidyl-peptidase IV activity and their association withother laboratory parameters. Clin Chem Lab Med 2001;39:155–9.

[20] Abbott CA, Yu DM, Woollatt E, Sutherland GR, McCaughan GW, Gorrell MD. Clon-ing, expression and chromosomal localization of a novel human dipeptidyl pepti-dase (DPP) IV homolog, DPP8. Eur J Biochem 2000;267:6140–50.

[21] Olsen C, Wagtmann N. Identification and characterization of human DPP9, a novelhomologue of dipeptidyl peptidase IV. Gene 2002;299:185–93.

[22] Bland JM, Altman DG. Statistical methods for assessing agreement between twomethods of clinical measurement. Lancet 1986;1:307–10.

[23] Lankas GR, Leiting B, Roy RS, et al. Dipeptidyl peptidase IV inhibition for the treat-ment of type 2 diabetes: potential importance of selectivity over dipeptidyl pep-tidases 8 and 9. Diabetes 2005;54:2988–94.

[24] Matheeussen V, Dubois V, Van Goethem S, et al. Application of clinically usedDPPIV-inhibitors in experimental in vivo and in vitro settings. Drugs of the future,35; 2010. p. 261.

[25] Beconi MG, Reed JR, Teffera Y, et al. Disposition of the dipeptidyl peptidase 4 in-hibitor sitagliptin in rats and dogs. Drug Metab Dispos 2007;35:525–32.

[26] D'Amico M, Di Filippo C, Marfella R, et al. Long-term inhibition of dipeptidylpeptidase-4 in Alzheimer's prone mice. Exp Gerontol 2010;45:202–7.

[27] Duttaroy A, Voelker F, Merriam K, et al. The DPP-4 inhibitor vildagliptin increasespancreatic beta cell mass in neonatal rats. Eur J Pharmacol 2011;650:703–7.

[28] Ferreira L, Teixeira-de-Lemos E, Pinto F, et al. Effects of sitagliptin treatment ondysmetabolism, inflammation, and oxidative stress in an animal model of type 2diabetes (ZDF rat). Mediators Inflamm 2010;2010:592760.

[29] Kawashima S, Matsuoka TA, Kaneto H, et al. Effect of alogliptin, pioglitazone andglargine on pancreatic beta-cells in diabetic db/db mice. Biochem Biophys ResCommun 2011;404:534–40.

[30] Maiztegui B, Borelli MI, Madrid VG, et al. Sitagliptin prevents the development ofmetabolic and hormonal disturbances, increased beta-cell apoptosis and liversteatosis induced by a fructose-rich diet in normal rats. Clin Sci (Lond)2011;120:73–80.

[31] Matsui T, Nishino Y, Takeuchi M, Yamagishi S. Vildagliptin blocks vascular injuryin thoracic aorta of diabetic rats by suppressing advanced glycation endproduct-receptor axis. Pharmacol Res 2011;63:383–8.

462 V. Matheeussen et al. / Clinica Chimica Acta 413 (2012) 456–462

[32] Matveyenko AV, Dry S, Cox HI, et al. Beneficial endocrine but adverse exocrine ef-fects of sitagliptin in the human islet amyloid polypeptide transgenic rat model oftype 2 diabetes: interactions with metformin. Diabetes 2009;58:1604–15.

[33] Mu J, Petrov A, Eiermann GJ, et al. Inhibition of DPP-4 with sitagliptin improvesglycemic control and restores islet cell mass and function in a rodent model oftype 2 diabetes. Eur J Pharmacol 2009;623:148–54.

[34] Tofovic DS, Bilan VP, Jackson EK. Sitagliptin augments angiotensin II-inducedrenal vasoconstriction in kidneys from rats with metabolic syndrome. Clin ExpPharmacol Physiol 2010;37:689–91.

[35] Vaghasiya J, Sheth N, Bhalodia Y, Manek R. Sitagliptin protects renal ischemia re-perfusion induced renal damage in diabetes. Regul Pept 2011;166:48–54.

[36] Kim SJ, Nian C, McIntosh CH. Sitagliptin (MK0431) inhibition of dipeptidyl pepti-dase IV decreases nonobese diabetic mouse CD4+ T-cell migration throughincretin-dependent and -independent pathways. Diabetes 2010;59:1739–50.

[37] Lamont BJ, Drucker DJ. Differential antidiabetic efficacy of incretin agonists versusDPP-4 inhibition in high fat fed mice. Diabetes 2008;57:190–8.

[38] Maida A, Hansotia T, Longuet C, Seino Y, Drucker DJ. Differential importance ofglucose-dependent insulinotropic polypeptide vs glucagon-like peptide 1 recep-tor signaling for beta cell survival in mice. Gastroenterology 2009;137:2146–57.

[39] Migoya EM, Bergeron R, Miller JL, et al. Dipeptidyl peptidase-4 inhibitors admin-istered in combination with metformin result in an additive increase in the plas-ma concentration of active GLP-1. Clin Pharmacol Ther 2010;88:801–8.

[40] Raun K, von Voss P, Gotfredsen CF, Golozoubova V, Rolin B, Knudsen LB. Liraglu-tide, a long-acting glucagon-like peptide-1 analog, reduces body weight andfood intake in obese candy-fed rats, whereas a dipeptidyl peptidase-IV inhibitor,vildagliptin, does not. Diabetes 2007;56:8–15.

[41] Bergman AJ, Stevens C, Zhou Y, et al. Pharmacokinetic and pharmacodynamicproperties of multiple oral doses of sitagliptin, a dipeptidyl peptidase-IV inhibitor:a double-blind, randomized, placebo-controlled study in healthy male volunteers.Clin Ther 2006;28:55–72.