Clinical Chemistry 56:2 272–280 (2010)
Proteomics and Protein Markers
Endoprotease Profiling with Double-Tagged Peptide Substrates: A New Diagnostic Approach in Oncology Teresa Peccerella,1 Nadine Lukan,2 Ralf Hofheinz,2 Dirk Schadendorf,3 Markus Kostrezewa,4 Michael Neumaier,1 and Peter Findeisen1*
BACKGROUND: The measurement of disease-related proteolytic activity in complex biological matrices like serum is of emerging interest to improve the diagnosis of malignant diseases. We developed a mass spectrometry (MS)-based functional proteomic profiling approach that tracks degradation of artificial endoprotease substrates in serum specimens. METHODS:
The synthetic reporter peptides that are cleaved by tumor-associated endopeptidases were systematically optimized with regard to flanking affinity tags, linkers, and stabilizing elements. Serum specimens were incubated with reporter peptides under standardized conditions and the peptides subsequently extracted with affinity chromatography before MS. In a pilot study an optimized reporter peptide with the cleavage motif WKPYDAADL was added to serum specimens from colorectal tumor patients (n ⫽ 50) and healthy controls (n ⫽ 50). This reporter peptide comprised a known cleavage site for the cysteineendopeptidase “cancer procoagulant.”
RESULTS:
Serial affinity chromatography using biotinand 6xHis tags was superior to the single affinity enrichment using only 6xHis tags. Furthermore, protease-resistant stop elements ensured signal accumulation after prolonged incubation. In contrast, signals from reporter peptides without stop elements vanished completely after prolonged incubation owing to their total degradation. Reporter-peptide spiking showed good reproducibility, and the difference in proteolytic activity between serum specimens from cancer patients and controls was highly significant (P ⬍ 0.001).
CONCLUSIONS: The introduction of a few structural key elements (affinity tags, linkers, d-amino acids) into
1
Institute for Clinical Chemistry, Medical Faculty Mannheim of the University of Heidelberg, University Hospital Mannheim, Mannheim, Germany; 2 III. Medical Clinic, Medical Faculty Mannheim of the University of Heidelberg, University Hospital Mannheim, Mannheim, Germany; 3 Skin Cancer Unit, German Cancer Research Center Heidelberg, and Department of Dermatology, University Hospital of Mannheim, Mannheim, Germany; 4 Bruker Daltonik, Bremen, Germany. * Address correspondence to this author at: Klinikum Mannheim, Institut fu¨r
272
synthetic reporter peptides increases the diagnostic sensitivity for MS-based protease profiling of serum specimens. This new approach might lead to functional MS-based protease profiling for improved disease classification. © 2009 American Association for Clinical Chemistry
High-throughput mass spectrometry (MS)5 can generate peptide patterns from hundreds of specimens that can be further analyzed with bioinformatics algorithms to enable early diagnosis and improve therapeutic monitoring of patients (1 ). However, the clinical proteomic profiling approach is fraught with controversy (2 ). For the analysis of serum specimens preanalytical variables have a major impact on profiling experiments by partially overriding disease-related peptide patterns (3 ) or even completely abolishing meaningful data interpretation (4 ). Furthermore, there are inherent difficulties in reproducing results from different studies (5, 6 ), and therefore clinical proteomic profiling has not matured to the point of being introduced into routine diagnostic procedures. However, under strictly controlled conditions, the proteolytic fragments from highly abundant endogenous serum proteins generate distinct peptide patterns that are suitable as surrogate marker for disease-specific proteolytic activity (7 ). This approach can further be improved through the use of exogenous reporter peptides that are added to serum specimens (8 ). The reaction conditions for the proteolytic degradation of reporter peptides can easily be controlled and standardized (8 –10 ) and the MS read out can be optimized accordingly (8 ). This approach is very similar to established diagnostic assays measuring the proteolytic activity of distinct enzymes, e.g., coagulation factors (11 ), and alleviates reproduc-
Klinische Chemie, Mannheim 68167, Germany. Fax ⫹49-(0)-621-383-3819; e-mail
[email protected]. Received July 29, 2009; accepted October 30, 2009. Previously published online at DOI: 10.1373/clinchem.2009.133462 5 Nonstandard abbreviations: MS, mass spectrometry; QCT, QC serum sample from 1 colorectal tumor patient; QCT, QC serum sample from 1 healthy control individual.
Protease Profiling with Reporter Peptides and Mass Spectrometry
Table 1. Different generation reporter peptides. The composition of amino- and carboxy-terminal modifications, as well as the introduction of distinct linkers, characterizes the different generations of the reporter peptides. Reporter peptide generation
N-terminal affinity tag
Linker
Biotin
III
Biotin
Abu
III
Ahx
ERGFFYTP
HHHHHHb
Ahx
ERGFFYTP
b
Ahx
Abu
Ahx
WKPYDAADL
HHHHHH
D-aaa
Ahx
HHHHHH
m/z [MⴙH]ⴙ
1838.83 2177.44 a
HHHHHHb
WKPYDAADL Biotin
C-terminal affinity tag
HHHHHHb
I
a
Linker
ERGFFYTP
I II
Amino acid sequence
2448.59 1900.87
t
2540.64
D-aa, D-amino acid. Amidation of C-terminus.
ibility problems related to sample collection, storage, and handling, as well as possible variability in endogenous peptide precursor concentrations (12 ). Finally, these assays would allow amplification of the output signals, potentially visualizing activity of lowabundant proteases in a complex proteome background (8 ). The success of functional endoprotease profiling with reporter peptide spiking has 3 essential prerequisites. First, the fragments of cleaved reporter peptides must be stabilized because of the multiple proteases acting within serum specimens. The initial cleavage of reporter peptides by endoproteases is followed by further proteolytic degradation through exoproteases (13, 14 ). This process results in the loss of MS signal intensity for transient fragments of reporter peptides when the incubation time is prolonged. Second, because serum is a very complex matrix, fractionation strategies, such as affinity chromatography (15 ), are essential to reduce unwanted ion suppression in MALDITOF MS analysis (16 ). Third, suitable reporter peptides with specific amino acid sequences that are selectively cleaved by tumor-associated proteases have to be identified in a systematic substrate screen (17, 18 ). Our goal was to optimize the first 2 prerequisites for protease profiling with reporter peptides to maximize the analytical sensitivity for the MS detection of cleaved fragments. For assay optimization the reporter peptides were synthesized with various affinity tags, linkers, and stabilizing elements. Finally, an optimized reporter peptide was added to serum specimens from colorectal cancer patients and healthy controls. Materials and Methods REPORTER PEPTIDES
The substrates used in experiments were peptides synthesized by the peptide synthesis facility of the German Cancer Research Center (Heidelberg, Germany). The substrates and their sequence information are shown
in Table 1. All peptides were prepared as a 4 mmol/L stock solution in DMSO. The sequences were derived from a literature search and the Merops peptidase database (19 ). The cleavage motif ERGFFYTP is derived from the insulin  chain (P01317) and is known to be cleaved by the tumor-associated protease cathepsin D (A01.009) (20 ). The cleavage motif WKPYDAADL is derived from the coagulation factor X heavy chain (P00742) and is known to be cleaved by the tumorassociated protease cancer procoagulant (EC 3.4.22.26) (21 ). SERUM SPECIMENS
Whole blood specimens were obtained from colorectal cancer patients and healthy controls at the University Hospital Mannheim. Blood collection was performed after we obtained institutional review board approval and patients’ written informed consent. The 50 latestage colorectal tumor patients included 25 men (median age 70 years, range 49 – 88 years) and 25 women (median age 63 years, range 47– 84 years). The 50 healthy controls included 25 men (median age 55 years, range 47– 65 years) and 25 women (median age 58 years, range 54 –73 years). We collected 10 mL of whole blood in plain tubes (Sarstedt). After a 30 min clot time at room temperature the specimens were centrifuged at 20 °C for 10 min at 3000g, then the serum was removed, divided into aliquots, and stored at – 80 °C until further use. REPORTER PEPTIDE SPIKING
Substrates were diluted to 100 mol/L in PBS pH 7.4 (PAA Laboratories). Each serum sample was diluted 1:3 in PBS. For the proteolytic profiling, 50 L of diluted serum sample was mixed with 50 L substrate solution and incubated at 37 °C for 2 h. During the preliminary experiments for the optimization of the protocol, incubation times ranging from 1 h to 16 h were used (as specified in the corresponding figures). Clinical Chemistry 56:2 (2010) 273
PURIFICATION OF PROTEOLYTIC FRAGMENTS
For purification of the 6xHis-tagged peptides, 20 L of diluted serum that had been incubated with substrate for 2 h at 37 °C was mixed with 60 L binding buffer and 10 L IMAC SL (immobilized metal ion affinity chromatography on self-loading) magnetic beads (Bruker Daltonics) that were loaded with Ni2⫹ ions, and then further processed according to manufacturer’s instructions. Briefly, the samples were incubated for 30 min at room temperature, washed 2 times each with 3 different buffers (100 L per wash step), and eluted with 10 L elution buffer for 15 min at room temperature. For the serial purification, streptavidin Sepharose high-performance slurry (GE Healthcare) was washed twice and resuspended in PBS pH 7.4 to yield a 50% slurry. We then added 50 L streptavidin slurry to each sample and incubated the mixture for 15 min at room temperature. After binding, the streptavidin Sepharose was centrifuged at 1100g for 5 min. A 20-L aliquot of the supernatant was further purified using the IMAC SL nickel magnetic beads as described above. A scheme of the work flow for the serial purification of reporter peptides is shown in Fig. 1. Theaminoterminalbiotintagofthesecond-generation reporter peptide Biotin-Ahx-ERGFFYPHHHHHH had to be protected from enzymatic degradation by Biotinidase (EC 3.5.1.12), which is inhibited with iodacetamide (22 ). We performed serial dilution experiments to further compare the efficiency of these 2 protocols for extracting proteolytic fragments in the presence of fulllength reporter peptides. Mixing 50 L of 200 mol/L substrate Biotin-Abu-Ahx-WKPYDAADLAhx-HHHHHHt (Table 1) and 50 L pooled serum from cancer patients resulted in a final substrate concentration of 100 mol/L. This was incubated overnight at 37 °C to ensure that the substrate was completely processed. The complete 100-L mixture was then subjected to serial purification, and eluted in 50 L to end up with the original volume of the substrate solution. This generated a solution containing only the anchor peptide Ahx-HHHHHHt with an m/z of 1055.1. The 50 L was then used for the serial dilution toward the detection limit in a background of 100 mol/L full-length peptide substrate. For dilution steps heat-inactivated serum (20 min at 65 °C) of a healthy control individual was used. To extract the reporter peptides the single purification protocol and the serial purification protocol were performed in parallel.
onto a SCOUT 600-m AnchorChip target (Bruker Daltonics). MALDI-TOF MS measurements were performed using an Autoflex II (Bruker Daltonics) operating in the positive mode. The matrix solution was prepared by dissolving 0.3 g/L of ␣-cyano-4hydroxycinnamic acid solution (Bruker Daltonics) in ethanol/acetone (2:1, vol/vol), and external mass calibration was then performed. The MS spectra for peaks in the range of 1–3 kDa were generated by summarizing 500 laser shots (50 laser shots at 10 different spot positions). For each spectrum a peak-picking algorithm (centroid, signal-to-noise ratio ⬎10) was applied by using flexAnalysis software (Bruker Daltonics).
MALDI-TOF ANALYSIS
CLINICAL PROTEASE PROFILING
We diluted 1 L of each eluted sample 1:10 in matrix solution, and 1 L of the resulting mixture was spotted
In a pilot study an optimized reporter peptide with the cleavage site WKPYDAADL was added to serum spec-
274 Clinical Chemistry 56:2 (2010)
Fig. 1. Work flow of the peptide extraction. Third-generation reporter peptides are added to serum specimens. After incubation for 2 h at 37 °C the unprocessed reporter peptides and proteolytic fragments that are labelled with biotin are depleted with streptavidin sepharose (I). The remaining his-tag labelled anchorpeptides are extracted with Ni-chelate beads (II). After elution (III) the peptides are analyzed with MALDI-TOF MS (IV). XX, linker; D-aa, D-amino acid.
Protease Profiling with Reporter Peptides and Mass Spectrometry
imens of colorectal tumor patients and healthy controls. This sequence is derived from the coagulation factor X (P00742) that is known to be cleaved by cancer procoagulant (19 ), a cysteine protease associated with many solid tumors, including colorectal cancer (23 ). The serum specimens from the colorectal cancer patients and healthy controls were strictly randomized before further sample handling, processing, and analysis. To monitor the reproducibility of reporter peptide spiking, 2 distinct QC samples were constructed that comprised serum specimens from 1 colorectal tumor patient (QCT) and 1 healthy control individual (QCH). Both samples were divided into aliquots and stored at ⫺80 °C until further use. On 4 consecutive days a set of 35 specimens comprising 25 randomized serum samples from tumor patients and healthy controls, plus 5 QCT samples and 5 QCH samples, were spiked with the reporter peptide (Biotin-Abu-AhxWKPYDAADL-Ahx-HHHHHHt) as described above. After incubation for 2 h at 37 °C the 6xHis tagged C-terminus of the cleaved reporter peptide was extracted with serial affinity chromatography as described above. The proteolytic processing of the reporter peptide Biotin-Abu-Ahx-WKPYDAADL-Ahx-HHHHHHt results in the accumulation of the anchor peptide AhxHHHHHHt (m/z 1055.1), and the respective peak intensities were used for quantification (15 ). The distribution of the m/z 1055.1 peak was visualized by use of whisker-box plots, and respective ROC calculation was performed with MedCalcTM software, version 8.0. The anchor peptide (m/z 1055.1) was fragmented for further verification with MALDI-TOF tandem MS (see Fig. 1 in the Data Supplement that accompanies the online version of this article at http:// www.clinchem.org/content/vol56/issue2) on an Autoflex II with LIFT upgrade (Bruker Daltonics). Results The first reporter peptide sequence ERGFFYTP is derived from oxidized insulin B chain (P01317) and is cleaved by cathepsin D (EC 3.4.23.5) (19 ) that is overexpressed in tumor tissue of colorectal cancer patients and furthermore is increased in serum specimens of tumor patients (20 ). The cleavage motif was modified with a carboxy terminal 6xHis-tag (Table 1) to enable the selective extraction of cleaved reporter peptide fragments from serum specimens (15 ). The sequence ERGFFYTPHHHHHH was designated as the “firstgeneration” reporter peptide. The pooled serum from colorectal tumor patients was spiked with the reporter peptide, and after 1 h incubation and subsequent affinity chromatography enrichment with Ni-chelate before MALDI-TOF MS a
peak with m/z 1349.4 was present (Fig. 2). This peptide corresponded to the fragment FYTPHHHHHH predicted to be cleaved from the reporter peptide ERGFFYTPHHHHHH by cathepsin D (19 ). The second reporter peptide sequence WKPYDAADL is part of the coagulation factor X heavy chain (P00742). The cysteine protease cancer procoagulant (EC 3.4.22.26) is secreted from various tumors including colorectal cancer (24 ) and cleaves the sequence WKPYDAADL (19 ). Again, after 1 h incubation with pooled serum of colorectal cancer patients and subsequent affinity chromatography enrichment with Nichelate before MALDI-TOF MS, a specific peak with m/z 1325.4 was present (see online Supplemental Fig. 2 and Supplemental Table 1) that corresponded to the predicted fragment DAADLHHHHHH (19 ). However, many other 6xHis tagged fragments of the reporter peptides were present after 1 h incubation, and the reporter peptide is trimmed down by exoproteases that generate a ladder-like degradation pattern (see online Supplemental Table 2) (13, 14 ). Most important, the signals vanished or decreased with prolonged incubation time (Fig. 2 and online Supplemental Fig. 2). To protect the reporter peptides from exoproteases, the amino terminus was blocked with a biotin-tag (25 ). The amidation of the carboxy terminus has been reported to protect against carboxypeptidases (26 ), and exoprotease resistance was characteristic to designated “second-generation” reporter peptides (Table 1). Interestingly, a mass shift of m/z 226 was observed as soon as second-generation reporter peptides were incubated with serum specimens (see online Supplemental Fig. 3). This mass shift corresponded to the loss of the biotin-tag cleaved off enzymatically by biotinidase (EC 3.5.1.12) (27 ) but can be stabilized by the addition of iodacetamide (22 ). Online Supplemental Fig. 4A shows the extended incubation for up to 16 h of the second-generation reporter peptide Biotin-AhxERGFFYTPHHHHHH with pooled serum from colorectal cancer patients. Although we introduced stabilizing elements, specific fragments (m/z 1038.5 and m/z 1201.6) of the reporter peptide decreased after 6 h and vanished after 16 h incubation (see online Supplemental Fig. 4A), results that were rather disappointing. We concluded that not only the reporter peptide but also resulting proteolytic fragments must be protected from exoproteases. The designated “third-generation” reporter peptides include additional elements to stabilize the 6xHistag (Table 1). A 6-aminohexanoic acid linker was introduced to stabilize the N-terminus of the 6xHis-tag. Furthermore, amidation of the C-terminus is not sufficient to suppress any carboxypeptidase activity (28 ). Instead, a D-amino acid (29 ) was introduced at the C-terminus to create a stable anchor peptide. This staClinical Chemistry 56:2 (2010) 275
Fig. 2. Proteolytic degradation of the first-generation reporter peptide ERGFFYTPHHHHHH. The first-generation reporter peptide ERGFFYTPHHHHHH (100 mol/L) was incubated with pooled serum from colorectal cancer patients in a time series as indicated in the graph. The 6xHis-tagged peptides were extracted with Ni chelate prior to MALDI-TOF MS. The following peaks correlate to proteolytic fragments: m/z 1838.8, ERGFFYTPHHHHHH; m/z 1553.3, RGFFYTPHHHHHH; m/z 1496.2, FFYTPHHHHHH; m/z 1349.0, FYTPHHHHHH; m/z 1201.8, YTPHHHHHH; m/z 1038.7, TPHHHHHH. Minor peaks with a mass shift of ⫹22 Da that originate from single-sodium adduct ions [M ⫹ Na]⫹ are indicated by ⫹. Signal intensity is in arbitrary units (au).
bilized 6xHis tag did not degrade after initial cleavage of third-generation reporter peptides by endopeptidases, but instead accumulated with prolonged incubation time (see online Supplemental Fig. 4B). Finally, the biotin-tag was protected from serum biotinidase activity by incorporation of 2-aminobutyric acid (27, 30 ) to avoid the use of iodacetamid, which inhibits not only biotinidase but also cysteine proteases (31 ). The protocol was optimized to extract fragments of cleaved reporter peptides in the presence of an excess of unprocessed full-length reporter peptides from serum specimens to improve the limit of detection (Fig. 3). Anchor peptides were either extracted with Ni2⫹loaded magnetic beads alone or with serial purification using streptavidin Sepharose and Ni2⫹-loaded magnetic beads in combination. With serial purification, the first step almost completely depleted any fulllength peptide, as can be seen in Fig. 3A, thereby facilitating the detection of low-abundant proteolytic fragments. The signal intensity of the anchor peptide was approximately 4-fold higher with the serial extraction 276 Clinical Chemistry 56:2 (2010)
protocol compared to the single affinity purification protocol (Fig. 3B). For the final step we set up a proof-of-principle experiment with the reporter peptide Biotin-Abu-AhxWKPYDAADL-Ahx-HHHHHHt to elucidate the applicability of protease profiling for diagnostic purposes. However, the implementation of MS as a routine diagnostic tool clearly depends on good reproducibility of the method. Therefore, in repetitive experiments we determined the reproducibility of reporter peptide spiking and subsequent serial extraction of the anchor peptide Ahx-HHHHHHt m/z 1055.1 before MALDITOF MS. QC samples QCT and QCH were randomly integrated into small series of serum specimens from colorectal cancer patients and control individuals over 4 consecutive days. Reporter peptide spiking of serum specimens, incubation, serial extraction of the anchor peptide (m/z 1055.1), and MALDI-TOF MS analysis were performed under strictly standardized conditions with respect to reporter peptide concentration, incubation time, temperature, and extraction of reporter peptide fragments before MALDI-TOF MS.
Protease Profiling with Reporter Peptides and Mass Spectrometry
Fig. 3. Dilution series of third-generation anchor peptides. (A), Decreasing concentrations of anchor peptide Ahx-HHHHHHt (m/z 1055.1) were diluted in 50 mol/L of full-length reporter peptide Biotin-Abu-Ahx-WKPYDAADL-Ahx-HHHHHHt (m/z 2540.2). The background matrix was heat inactivated serum of a healthy donor. Panel I, his-tag purification only; panel II, serial purification. (B), Zoom of the bottom panel corresponding to 1.5 mol/L anchor peptide (m/z 1055.1) in an excess background of 50 mol/L full-length reporter peptide (m/z 2540). Panel I, his-tag purification only; panel II, serial purification. au, arbitrary units.
To test the reproducibility of reporter peptide spiking, we analyzed 2 QC samples (QCT, QCH) 5 times a day on 4 consecutive days. The reporter peptide Biotin-Abu-Ahx-WKPYDAADL-Ahx-HHHHHHt was
added to aliquots of the QCT and QCH serum, which were incubated, and then the anchor peptide (m/z 1055.1) was serially extracted. The CVs for intraday reproducibility of the QCT sample ranged from 26.5% Clinical Chemistry 56:2 (2010) 277
Fig. 4. Reproducibility of reporter-peptide spiking. We incubated 100 mol/L of full-length reporter peptide Biotin-Abu-Ahx-WKPYDAADL-Ahx-HHHHHHt (m/z 2540.2) with serum specimens as described in Materials and Methods. The anchor peptide Ahx-HHHHHHt (m/z 1055.1) was extracted with serial affinity chromatography prior to MALDI-TOF MS. Signal intensity of the peak m/z 1055.1 is shown. Light grey, QCT; dark grey, QCH. au, arbitrary units.
(mean 47.9, SD 12.7) to 8.4% (mean 34.4, SD 2.9) and CVs of the QCH sample ranged from 55.2% (mean 5.0, SD 2.8) to 16.6% (mean 1.6, SD 1.2). The CV for interday reproducibility was 32% (mean 32.2, SD 9.9) for the QCT sample and 37% (mean 9.9, SD 3.6) for the QCH sample. The QCT sample could clearly be separated from QCH samples at any day (Fig. 4). In a small profiling experiment (Fig. 5), serum specimens of colorectal cancer patients were compared to those of healthy controls. The signal intensities of the m/z 1055.1 anchor peptide were significantly higher in serum specimens from colorectal cancer patients (Fig. 5A). The median signal intensity of the anchor pepetide m/z 1055.1 in colorectal cancer serum specimens was 84530, with a low value of 3466 and high value of 127 461. The median signal intensity of the anchor peptide m/z 1055.1 in serum specimens of healthy control individuals was 19 364 with a low value of 2317 and high value of 122 351 (Fig. 5A). The area under the ROC-curve was 0.8 (Fig. 5B). Discussion Most serum biomarkers that have been identified from MS-based clinical proteomic profiling studies are only proteolytic fragments from relatively abundant serum proteins (13, 14 ). However, the distinct fragment patterns of high-abundant serum proteins have been demonstrated to be useful as potential biomarkers for detection and classification of cancer (14 ). These patterns originate from proteolytic processing of high278 Clinical Chemistry 56:2 (2010)
Fig. 5. Small profiling experiment. (A), Proteolytic processing of a reporter peptide in serum specimens from colorectal cancer patients and healthy controls. The horizontal bars of the box plot indicate the 95% CI for the median. (B), ROC curve for the anchor peptide (m/z 1055.1). Area under the curve is 0.8. au, arbitrary units.
abundant proteins ex vivo by a multitude of proteases that include endopeptidases and exopeptidases (14 ). Proteases are known to play a crucial role in malignant disease, and tumor-specific proteases are released into the bloodstream during tumor progression and metastasis (32–34 ). Recent reports demonstrated that spiking of exogenous reporter peptides into serum for characterization of protease activity offers substantial advantages over profiling of native serum with respect to improved standardization (8 –10 ). This approach, however, is restricted to the monitoring of exoproteases activity (8 ) and cannot detect many tumor-associated endoproteases. To circumvent this serious shortcoming our study aimed at generating substrates for the more fastidious endoproteases. The stability of peptides in human serum and plasma is lim-
Protease Profiling with Reporter Peptides and Mass Spectrometry
ited owing to generically high activity of proteases in these biological specimens (35 ). To protect peptides from unspecific and unwanted proteolytic processing, distinct modifications must be introduced into the amino acid sequence (36 ). The carboxy and amino termini of the reporter peptides had to be protected from proteolytic processing by exoproteases. Furthermore, the fragments that resulted from proteolytic processing by endoproteases had to be stabilized to ensure their accumulation during prolonged incubation times. Finally, the peptide extraction protocol had to be optimized to detect low-abundant fragments of cleaved reporter peptides in an excess background of full-length reporter peptides that had been added to serum specimens. The excess of reporter peptides is necessary not only to ensure substrate saturation but also to displace competing natural substrates from the respective proteases. However, if protease activity is low, the excess of unprocessed full-length reporter peptides has to be depleted to avoid saturation of the affinity chromatography surface and ensure a sensitive detection of low-abundant proteolytic fragments. With a tandem affinity purification protocol, in a first step the biotinlabeled full-length reporter peptides were eliminated by binding to immobilized streptavidin. In a second step the remaining 6xHis tagged proteolytic fragments were enriched with Ni2⫹-loaded magnetic beads before MS. We simulated different levels of protease activity by spiking decreasing concentrations of a processed peptide fragment into a background of unprocessed full-length reporter peptide, e.g., enough activity to degrade 50% of substrate present down to only 3%. At the lowest concentration of anchor peptide, the double-tag approach still yielded a clear signal intensity that was overall 4 times higher than for the single-tag purification (Fig. 3). However, the reproducibility of the peptide extraction with consecutive MS analysis is a major issue if the approach is to be applicable for routine diagnostic testing. QC specimens were spiked with the reporter peptide, and the proteolytic fragment m/z 1055.1 was extracted on 4 consecutive days (Fig. 4). As expected, the CVs were higher for specimens with low signal intensity (QCH), ranging from 16.6% to 55.2%. In contrast, the CVs for specimens with high signal intensity (QCT) ranged from 8.4% to 26.5%, in accordance with a previous report (15 ). Finally, a small proof-of-principle protease profiling experiment was set up to elucidate the classification power of reporter peptide spiking to differentiate colorectal cancer patients from healthy control individuals. The difference of median signal intensity for the anchor peptide m/z 1055.1 between the 2 collectives was highly different (Fig. 5A). However, few outliers of the healthy control individuals showed high signal activity for the peak m/z 1055.1, hypothesized to be a surrogate marker for can-
cer procoagulant activity. This, in part, might be due to the fact, that cancer procoagulant is not exclusively present in serum specimens from tumor patients but also can be found in serum specimens from healthy control individuals (37 ). It might also be possible that the reporter peptide is not only cleaved by cancer procoagulant but also by other endoproteases, resulting in decreased diagnostic specificity of the test. In contrast, some cancer patients had low signal intensities for the m/z 1055.1 peak, indicating no enhanced activity of cancer procoagulant (Fig. 5A) and leading to decreased diagnostic sensitivity of the test. We emphasize that this small proof-of-principle profiling experiment has serious shortcomings concerning the limited number of analyzed specimens and the comparison of healthy controls with late-stage tumor patients, which is not relevant from a clinical point of view. Therefore, the biomarker we describe here is not thoroughly validated according to recent recommendations (38 ), but is thought to illustrate the feasibility of functional protease profiling. It is likely that tumor heterogeneity and progression of malignant disease result in different protease patterns (39, 40 ). The use of one single reporter peptide seems not to be sufficient to ensure high diagnostic accuracy and accordingly the area under the ROC curve was only 0.8 in our profiling experiment (Fig. 5B). To increase the diagnostic sensitivity and specificity of functional protease profiling with reporter peptides, it would seem reasonable to combine different reporter peptides. To achieve this goal, it will be necessary to systematically identify reporter peptide sequences that are most efficiently cleaved by disease-specific proteases (18 ). Furthermore, the characterization and identification of proteases involved in processing of reporter peptides is also essential to improve this approach. Most important, the addition of exogenous reporter peptides to serum offers major advantages concerning the standardization of preanalytical variabilities (9 ). Defining substrate concentration, incubation time, and temperature can effectively control the proteolytic reactions. In addition, the impact of sample collection and storage on the preservation of the endogenous protease activities that must be systematically elucidated must be taken into account.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Clinical Chemistry 56:2 (2010) 279
Authors’ Disclosures of Potential Conflicts of Interest: Upon manuscript submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest: Employment or Leadership: M. Kostrzewa, Bruker Daltonik. Consultant or Advisory Role: None declared. Stock Ownership: None declared. Honoraria: None declared. Research Funding: None declared.
Expert Testimony: None declared. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. Acknowledgments: We gratefully acknowledge that the costs of publication and reprints are being supported by the LESSER-LOEWE Foundation.
References 1. Pusch W, Flocco MT, Leung SM, Thiele H, Kostrzewa M. Mass spectrometry-based clinical proteomics. Pharmacogenomics 2003;4:463–76. 2. Diamandis E. Oncopeptidomics: a useful approach for cancer diagnosis? Clin Chem 2007;53:1004 – 6. 3. Baggerly K, Morris J, Edmonson S, Coombes K. Signal in noise: evaluating reported reproducibility of serum proteomic tests for ovarian cancer. J Natl Cancer Inst 2005;97:307–9. 4. Karsan A, Eigl BJ, Flibotte S, Gelmon K, Switzer P, Hassell P, et al. Analytical and preanalytical biases in serum proteomic pattern analysis for breast cancer diagnosis. Clin Chem 2005;51:1525– 8. 5. McLerran D, Grizzle WE, Feng Z, Thompson IM, Bigbee WL, Cazares LH, et al. SELDI-TOF MS whole serum proteomic profiling with IMAC surface does not reliably detect prostate cancer. Clin Chem 2008;54:53– 60. 6. McLerran D, Grizzle WE, Feng Z, Bigbee WL, Banez LL, Cazares LH, et al. Analytical validation of serum proteomic profiling for diagnosis of prostate cancer: sources of sample bias. Clin Chem 2008;54:44 –52. 7. Villanueva J, Shaffer DR, Philip J, Chaparro CA, Erdjument-Bromage H, Olshen AB, et al. Differential exoprotease activities confer tumor-specific serum peptidome patterns. J Clin Invest 2006;116:271– 84. 8. Villanueva J, Nazarian A, Lawlor K, Yi SS, Robbins RJ, Tempst P. A sequence-specific exopeptidase activity test (SSEAT) for “functional” biomarker discovery. Mol Cell Proteomics 2008;7:509 –18. 9. Findeisen P, Post S, Wenz F, Neumaier M. Addition of exogenous reporter peptides to serum samples before mass spectrometry-based protease profiling provides advantages over profiling of endogenous peptides. Clin Chem 2007;53:1864 – 6. 10. Findeisen P, Peccerella T, Post S, Wenz F, Neumaier M. Spiking of serum specimens with exogenous reporter peptides for mass spectrometry based protease profiling as diagnostic tool. Rapid Commun Mass Spectrom 2008;22:1223–9. 11. Witt I. Test systems with synthetic peptide substrates in haemostaseology. Eur J Clin Chem Clin Biochem 1991;29:355–74. 12. Nedelkov D, Kiernan UA, Niederkofler EE, Tubbs KA, Nelson RW. Investigating diversity in human plasma proteins. Proc Natl Acad Sci U S A 2005;102:10852–7. 13. Koomen J, Shih L, Coombes K, Li D, Xiao L-C, Fidler I, et al. Plasma protein profiling for diagnosis of pancreatic cancer reveals the presence of host response proteins. Clin Cancer Res 2005;11:1110 – 8. 14. Villanueva J, Martorella A, Lawlor K, Philip J, Fleisher M, Robbins R, Tempst P. Serum pep-
280 Clinical Chemistry 56:2 (2010)
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
tidome patterns that distinguish metastatic thyroid carcinoma from cancer-free controls are unbiased by gender and age. Mol Cell Proteomics 2006;5:1840 –52. Jin M, Cataland S, Bissell M, Wu HM. A rapid test for the diagnosis of thrombotic thrombocytopenic purpura using surface enhanced laser desorption/ ionization time-of-flight (SELDI-TOF)-mass spectrometry. J Thromb Thrombolysis 2006;4:333– 8. Schlosser G, Pocsfalvi G, Huszar E, Malorni A, Hudecz F. MALDI-TOF mass spectrometry of a combinatorial peptide library: effect of matrix composition on signal suppression. J Mass Spectrom 2005;40:1590 – 4. Xu D, Suenaga N, Edelmann M, Fridman R, Muschel R, Kessler B. Novel MMP-9 substrates in cancer cells revealed by a label-free quantitative proteomics approach. Mol Cell Proteomics 2008;7:2215–28. Turk BE, Cantley LC. Using peptide libraries to identify optimal cleavage motifs for proteolytic enzymes. Methods 2004;32:398 – 405. Rawlings ND, Morton FR, Kok CY, Kong J, Barrett AJ. MEROPS: the peptidase database. Nucleic Acids Res 2008;36:D320 –5. Szajda S, Snarska J, Jankowska A, RoszkowskaJakimiec W, Puchalski Z, Zwierz K. Cathepsin D and carcino-embryonic antigen in serum, urine and tissues of colon adenocarcinoma patients. Hepatogastroenterology 2008;55:388 –93. Kazmierczak M, Lewandowski K, Wojtukiewicz M, Turowiecka Z, Kolacz E, Lojko A, et al. Cancer procoagulant in patients with adenocarcinomas. Blood Coagul Fibrinolysis 2005;16:543–7. Craft DV, Goss NH, Chandramouli N, Wood HG. Purification of biotinidase from human plasma and its activity on biotinyl peptides. Biochemistry 1985;24:2471– 6. Gordon SG, Benson B. Analysis of serum cancer procoagulant activity and its possible use as a tumor marker. Thromb Res 1989;56:431– 40. Gordon SG, Mielicki WP. Cancer procoagulant: a factor X activator, tumor marker and growth factor from malignant tissue. Blood Coagul Fibrinolysis 1997;8:73– 86. Kozlov IA, Melnyk PC, Hachmann JP, Srinivasan A, Shults M, Zhao C, et al. A high-complexity, multiplexed solution-phase assay for profiling protease activity on microarrays. Comb Chem High Throughput Screening 2008;11:24 –35. Heavner G, Kroon D, Audhya T, Goldstein G. Biologically active analogs of thymopentin with enhanced enzymatic stability. Peptides 1986;7:1015–9. Wilbur D, Hamlin D, Pathare P, Weerawarna S. Biotin reagents for antibody pretargeting. Synthesis, radioiodination, and in vitro evaluation of
28.
29.
30.
31.
32.
33. 34.
35.
36.
37.
38.
39.
40.
water soluble, biotinidase resistant biotin derivatives. Bioconjug Chem 1997;8:572– 84. Malfroy B, Schwartz JC. Comparison of dipeptidyl carboxypeptidase and endopeptidase activities in the three enkephalin-hydrolysing metallopeptidases: “angiotensin-converting enzyme,” thermolysin and “enkephalinase.” Biochem Biophys Res Commun 1985;130:372– 8. Powell M, Stewart T, Otvos JL, Urge L, Gaeta F, Sette A, et al. Peptide stability in drug development. II. Effect of single amino acid substitution and glycosylation on peptide reactivity in human serum. Pharm Res 1993;10:1268 –73. Wilbur D, Hamlin D, Chyan M-K. Biotin reagents for antibody pretargeting, 7: investigation of chemically inert biotinidase blocking functionalities for synthetic utility. Bioconjug Chem 2006;17:1514 –22. Vasishta A, Baker P, Preece P, Wood R, Cuschieri A. Inhibition of proteinase-like peptidase activities in serum and tissue from breast cancer patients. Anticancer Res 1988;8:785–9. Ludwig T. Local proteolytic activity in tumor cell invasion and metastasis. BioEssays 2005;27: 1181–91. Garbett EA, Reed MW, Brown NJ. Proteolysis in colorectal cancer. Mol Pathol 1999;52:140 –5. Tang Y, Nakada M, Kesavan P, McCabe F, Millar H, Rafferty P, et al. Extracellular matrix metalloproteinase inducer stimulates tumor angiogenesis by elevating vascular endothelial cell growth factor and matrix metalloproteinases. Cancer Res 2005;65:3193–9. Powell M, Grey H, Gaeta F, Sette A, Colon S. Peptide stability in drug development: a comparison of peptide reactivity in different biological media. J Pharm Sci 1992;81:731–5. Walker J, Altman R, Warren J, Altman E. Using protein-based motifs to stabilize peptides. J Pept Res 2003;62:214 –26. Gordon SG, Cross BA. An enzyme-linked immunosorbent assay for cancer procoagulant and its potential as a new tumor marker. Cancer Res 1990;50:6229 –34. Mischak H. Clinical proteomics: A need to define the field and to begin to set adequate standards. Proteomics Clin Appl 2007;1:148 –56. Lin TS, Chiou SH, Wang LS, Huang HH, Chiang SF, Shih AY, et al. Expression spectra of matrix metalloproteinases in metastatic non-small cell lung cancer. Oncology Reports 2004;12:717–23. Bok RA, Hansell EJ, Nguyen TP, Greenberg NM, McKerrow JH, Shuman MA. Patterns of protease production during prostate cancer progression: proteomic evidence for cascades in a transgenic model. Prostate Cancer Prostatic Dis 2003;6:272– 80.
