WO2008143494A1

WO2008143494A1 - Assay for detection of prostate cancer by means of proteolytic hsa markers

Info

Publication number: WO2008143494A1
Application number: PCT/NL2007/050236
Authority: WO
Inventors: Theo Marten Luider; Guido Wijnandus Jenster
Original assignee: Erasmus University Medical Center Rotterdam
Priority date: 2007-05-22
Filing date: 2007-05-22
Publication date: 2008-11-27

Abstract

The present invention relates to a new assay for prostate cancer and peptide markers useful in said assay. It has been found that the serum of prostate cancer patients with metastasis comprises a different set of proteases and proteolytic enzymes than healthy controls. This results in different degradation products of human serum albumin (HSA), which can function as markers for assaying the presence of prostate cancer. Further, the proteolytic characteristics of the serum can be tested by applying a synthetic peptide to the serum and assaying for its degradation products. Specific proteolytic markers are peptides comprising the HSA fragments HPDYSWLL or HPYFYAPEL.

Description

Titlff

ASSAY FOR DETECTION OF PROSTATE CANCER BY MEANS OF PROTEOLYTIC HSA MARKERS

The invention is directed to a marker for the detection of prostate carcinoma, to methods for detecting a prostate carcinoma, and to a method for determining the efficacy of a prostate carcinoma treatment.

Prostate carcinoma is the most commonly diagnosed carcinoma in men over the age of 50. Cells in the prostate gland become abnormal and start to grow uncontrollably, forming tumours. A tumour in the prostate interferes with proper control of the bladder and normal sexual functioning. Carcinogenic cells within the prostate itself are generally not deadly on their own. However, as the tumour grows, some of the cells are released from the tumor and spread to other parts of the body through the lymph or the blood, a process known as metastasis.

Prostate carcinoma is curable when detected early. Yet the early stages of prostate carcinoma are often asymptomatic, so the disease often goes undetected until the patient has a routine physical examination. The currently used methods for diagnosing prostate carcinoma include screening for elevated prostate-specific antigen (PSA) levels, digital rectal examination, transrectal ultrasound imaging, and needle biopsy of the prostate.

During digital rectal examination, the rectum is examined for any lumps in the prostate. The rectum lies just behind the prostate gland, and a majority of prostate tumours begin in the posterior region of the prostate. Abnormalities may suggest the presence of prostate carcinoma.

In transrectal ultrasound a small sound wave releasing probe is placed in the rectum, allowing to create an image of reflected waves. Normal prostate tissue and prostate tumours reflect the sound waves differently. Therefore this test can be used to detect tumours. However, the insertion of the probe into the rectum is uncomfortable and this test is mainly carried out as confirmation if prostate carcinoma is already suspected. Prostate biopsy by removing a small piece of prostate tissue with a hollow needle is at present the most definitive diagnostic tool for prostate carcinoma. However, also this test is normally carried out as confirmation if carcinoma is already suspected from other tests. Screening of a patients blood for elevated PSA levels is in principle a simple procedure. The cells lining the prostate generally make PSA and a small amount can normally be detected in the bloodstream. In contrast, prostate carcinomas produce a large amount of this protein, significantly raising the circulating levels. A finding of a higher PSA level than normal with respect to the patient's age group therefore suggests the presence of prostate carcinoma.

However, the diagnosis of prostate carcinoma on the basis of PSA levels alone lacks sensitivity and specificity. Even in combination with other tests the sensitivity and specificity increase, but still remain poor and variable (S. Selley et al. (1997). Diagnosis, management and screening of early localised prostate carcinoma. Health Technol Assess 1, i, 1-96.) The normal range of PSA in serum is between 0-4 ng/ml and is subject to debate because varying proportions of men (25 %-73 %) with no evidence of prostate carcinoma have serum PSA levels above the upper limit of 4 ng/ml (W.J. Catalona et al. (1994). Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate carcinoma: results of a multicenter clinical trial of 6 630 men. J Urol 151, 1283-1290 and M.K. Brawer et al. (1993). Screening for prostatic carcinoma with prostate specific antigen: results of the second year. J Urol 150, 106-109). Conversely men with confirmed prostate carcinoma can have PSA levels below 4 ng/ml.

Accordingly, there is a need for new markers in a patient's body fluids for the detection of prostate carcinoma. Such new markers could be used in combination with the screening of PSA levels in order to reduce the number of false negative and false positive diagnosed patients. In addition, there is a need for prognostic markers which can be important in distinguishing patients that benefit from therapy from those that do not respond to therapy. Object of the present invention is to provide markers for the detection of prostate carcinoma.

Another object of the invention is to provide a method for detection and identification of prostate carcinoma specific proteolytic enzymes and inhibitors.

The inventors found that these objects are at least partly met by a specific peptide marker being a proteolytic degradation product of a serum sample.

Accordingly, in a first aspect the invention is directed to a peptide marker for the detection of prostate carcinoma, said peptide marker being a proteolytic degradation product of a serum, wherein said marker comprises an amino acid sequence derived from an amino acid sequence defined by SEQ ID No. 1 or 2.

Nowadays it is possible to perform protein/peptide profiling on large sample sets using different forms of mass spectrometry. Several studies have been published in which disease specific protein profiles are detected with high sensitivity and specificity (K.R. Kozak et al. (2005). Characterisation of serum biomarkers for detection of early stage ovarian cancer. Proteomics 5, 4589-4596; D. Sidransky et al. (2003). Serum protein MALDI profiling to distinguish upper aerodigestive tract cancer patients from control subjects. J Natl Cancer Inst 95, 1711-1717; E.F. Petricoin 3^rd et al, (2002). Serum proteomic patterns for detection of prostate cancer. J Natl Cancer Inst 94, 1576-1578; J. Villanueva et al. (2006). Serum peptidome patterns that distinguish metastatic thyroid carcinoma from cancer-free controls are unbiased by gender and age. MoI Cell Proteomics 5, 1840-1852). Most of the identified differentially expressed proteins are high abundant serum proteins or fragments thereof that are related to stress, infection, hormonal modulation and acute phase reactions (J.M. Koomen et al. (2005). Plasma protein profiling for diagnosis of pancreatic cancer reveals the presence of host response proteins. Clin Cancer Res 11, 1110-1118). The possible clinical applicability of such proteins or break-down products of proteins has been discussed (L.A. Liotta et al. (2003). Clinical proteomics: written in blood. Nature 425, 905). In addition some recent studies have shown that specific protease activities can be highly disease specific for different types of cancers (J. Villanueva et al. (2006). Differential exoprotease activities confer tumor-specific serum peptidome patterns. J Clin Invest 116, 271-284). The applicant has shown in a previous study that tryptic peptide profiling is a powerful technique to detect differentially expressed peptides (L.J. Dekker et al. (2005). MALDI-TOF mass spectrometry analysis of cerebrospinal fluid tryptic peptide profiles to diagnose leptomeningeal metastases in patients with breast cancer. MoI Cell Proteomics 4, 1341-1349).

The invention is based on the discovery that prostate carcinoma is accompanied by specific protease activities, which result in differentially expressed peptides. These specific protease activities and the resulting differentially expressed peptides can act as markers for the detection of prostate carcinoma and/or as indicators for evaluating the efficacy of a prostate carcinoma therapy.

A large number of abundant serum proteins are active in proteolysis or in the inhibition of proteolyses. The large number of proteolytic enzymes and inhibitors present in blood and a multi step proteolysis results in vast numbers of possible fragments. Degradation of proteins and peptides by proteolytic activity is a common phenomenon especially in metastasis of cancer (J.E. Koblinski et al. (2000). Unraveling the role of proteases in cancer. Clin Chim Acta 291, 113-135; J. Earn et al. (1998). Requirement for specific proteases in cancer cell intravasation as revealed by a novel semiquantitative PCR-based assay. Cell 94, 353-362; A.F. Chambers et al. (1997). Changing views of the role of matrix metalloproteinases in metastasis. J Natl Cancer Inst 89, 1260-1270). During metastasis these proteolytic enzymes are released in the blood and can possibly result also in an increased secretion of "specific" inhibitors (B. F. Sloane et al. (1992). Cysteine endopeptidases and their inhibitors in malignant progression of rat embryo fibroblasts. Biol Chem

Hoppe Seyler 373, 589-594). The variation in enzymes and specific inhibitors enables observation of disease specific protein/peptide profiles by mass spectrometry techniques.

In order to detect the differentially expressed peptides the inventors performed a tryptic peptide profiling experiment on tryptically digested serum samples as illustrated in the Example.

Specific mono-isotopic masses were found for which a clear difference in intensity was observed between control serum samples and prostate carcinoma with metastasis serum samples. Surprisingly, the identification of the mono-isotopic masses resulted in the identification of breakdown peptides from human serum albumin (HSA).

More in particular, it was found that the differentially expressed peptide markers of the invention are tryptic degradation fragments of two homologous amino acid sequences in the human serum albumin (HSA) protein defined by SEQ ID No. 3 and 4. Additional HSA fragments, defined herein as "semi- tryptic" fragments, i.e. fragments that have been degraded first by tryptic digestion and further degraded by other protease action that are derived from SEQ ID No. 3 or 4 and that are differentially expressed in serum samples from a subject suffering from prostate carcinoma are defined by SEQ ID No. 5-21. Preferred peptide markers, which are relatively abundant in prostate cancer patients are SEQ ID No. 7, 10, 15, 16 and 21, while the peptides of SEQ ID Nos. 5, 6, 8, 9, 11, 12, 13, 14, 17, 18, 19 and 20 are more abundant in healthy controls.

It should be noted that in accordance with the present invention the expression "a peptide marker comprising an amino acid sequence derived from an amino acid sequence of SEQ ID No. 3 or 4" is meant to refer to a peptide marker that comprises an amino acid sequence which is identical or substantially homologous with the amino acid sequence defined by SEQ ID No. 3 or 4. The term "substantially homologous" should be interpreted herein as two amino acid sequences having a "percentage of sequence homology" of at least 60, more preferably at least 80, even more preferably at least 90, still more preferably at least 95, still more preferably at least 98, and most preferably at least 99 percent amino acid sequence homology.

In addition, the peptide marker of the present invention can be a physiologically acceptable salts of the peptide defined by SEQ ID No. 1, 2 or 5-20.

As is shown in Figure 4, the peptides of SEQ ID Nos: 7, 10, 15, 16 and 21 are found predominantly in prostate cancer patients with metastasis, while the other tryptic and semi-tryptic peptides of Figure 3 are predominantly found in healthy controls. This means that the ratio of these peptides offers a useful discrimination between patients and controls and thus a useful diagnostic tool for the screening for prostate cancer (see also Table 3, where the ratio(s) are given in the last column).

The peptide markers of the invention can be identified from serum samples as described in the Example below by tryptic peptide profiling on tryptically digested serum samples. Assaying for the peptides in the tryptic digested serum samples is ideally performed by mass spectrography or any other means in which the length and/or mass of the tryptic and semi-tryptic digests can be determined. The ratio of the peptides of SEQ ID Nos: 7, 10, 15, 16 and 21 with any of the other peptides listed in Figure 4 is less than 1.0 in prostate cancer metastatic patients, while said ratio in greater than 1.0 in healthy controls.

More specifically, in a further aspect the present invention is directed to a method for detecting a prostate carcinoma comprising obtaining a first serum sample from a subject potentially suffering from prostate carcinoma and optionally a second serum sample from a subject not suffering from prostate carcinoma; subjecting said first serum sample and said optional second serum sample to tryptic digestion to obtain a first tryptic digest and a second tryptic digest, respectively; - analysing said first tryptic digest and said optional second tryptic digest for expressed peptides by peptide profiling; examining said first tryptic digest and said optional second tryptic digest for one or more differentially expressed peptides; and optionally comparing the results obtained with a control or panel of control samples; and assign a diagnosis on basis of the levels of the peptide fragments. According to the method of the invention a serum sample may be taken from a subject, which is potentially suffering from prostate carcinoma. The subject can be an animal, preferably a mammal, more preferably a human. Optionally, a second control serum sample is taken from a healthy subject, or a subject which is at least not suffering from prostate carcinoma. Said serum samples are then subjected to tryptic digestion during which the proteins present are hydrolysed into smaller polypeptide units.

In a next step the tryptic digests are analysed for expressed peptides using peptide profiling. Preferably this step is preceded by a magnetic bead purification step in which the peptides are bound to surface active magnetic beads. The beads to which the peptides are bound can be separated from the supernatant by a magnetic separation device. The bound and purified peptides can then be eluted from the beads. Such a step enriches the sample for proteins, while possibly disturbing salts and small molecules are discarded.

Subsequently, expressed peptides in the different tryptic digest sampels are examined by peptide profiling. The peptide profiling preferably is performed by mass spectrometry, more preferably it involves MALDI-TOF mass spectrometry, by which the peptides and peptide fragments are profiled according to there molecular weight, or electrospray mass spectrometry.

In order to overcome the problem of relatively poor reproducibility of peak intensities in MALDI-TOF techniques, which is mentioned in the prior art (K.A. Baggerly et al. (2004). Reproducibility of SELDI-TOF protein patterns in serum: comparing datasets from different experiments. Bioenformatics 20, 777-785), multiple replicates measurements per sample can be analysed. In addition, it is preferred that the peptide profiling is carried out more than once, preferably 1-10 times, more preferably 2-5 times, with time intervals of for instance 1-8 weeks, preferably 2-6 weeks, more preferably 4-5 weeks. In this way the stability of the expressed peptides over time can be studied.

The resulting MALDI-TOF mass spectra of the sample of the subject potentially suffering from prostate carcinoma and the sample of the subject not suffering from prostate carcinoma can be analysed and statistically compared.

As indicated above, a diagnosis that the sample is derived from a patient with (metastasis of) prostate cancer can be achieved by calculating the ratio between either peptide selected from the group of SEQ ID Nos: 7, 10, 15, 16 and 21 and either peptide selected from the group of SEQ ID Nos: 5, 6, 8, 9, 11, 12, 13, 14, 17, 18, 19 and 20. This ratio should be calculated between the peptides in the sample obtained from the subject potentially suffering from prostate carcinoma. The optional second sample of a healthy person can be used to validate the detection and calculation method. Alternatively, the optional second sample can be used to calibrate the levels of the first sample: the average level of the peptides from the group of SEQ ID Nos: 5, 6, 8, 9, 11, 12, 13, 14, 17, 18, 19 and 20 in the first sample should be recalculated to match the average level of said peptides in the second sample. Then, the level of the peptide selected from the group of SEQ ID Nos: 7, 10, 15, 16 and 21 should be adjusted by multiplying with the same factor, after which the ratio(s) between any of the peptides selected from the group of SEQ ID Nos: 7, 10, 15, 16 and 21 from the first sample and the corresponding peptide selected from the group of SEQ ID Nos: 7, 10, 15, 16 and 21 in the second sample can be calculated. The sample is positive (i.e. the subject is suffering from prostate cancer) if said ratio is larger than 1.0. It is thus possible to assay the peptide markers of the invention by determining the mass and/or length of the identified peptides in a tryptically digested serum sample.

An other preferred assay of the invention is performed by using synthetically prepared peptide fragments to assay the protease activity in serum samples. The person skilled in the art is familiar with conventional peptide synthesis procedures and is able to construct the synthetic peptides discussed below.

As shown by the inventors, the specific protease activity for proteolytic degradation of HSA fragments is highly specific for prostate cancer. In another embodiment, therefore, the invention is directed to a method for detecting a prostate carcinoma comprising obtaining a serum sample from a subject potentially suffering from prostate carcinoma and optionally a control sample (such as a serum sample from a subject not suffering from prostate carcinoma); - measuring the specific protease activity by adding a peptide having the sequence of SEQ ID No:7 or 10 analysing the peptide content of the sample after the protease reaction (if any).

Of course, the degradation of the peptide increases with the time of the assay and decreases with dilution of the serum. It is therefore important to calibrate the assay by performing a time series and/or a dilution series, and to choose a time and dilution step for the assay which allow for an optimal difference between the peptide degradation products resulting from the protease reaction. In general, about two hours incubation of the synthetic peptide with about a 100-fold diluted serum would yield useful results.

It has appeared that in assays as described above the same degradation pattern of the synthetic peptides used is seen as in tryptic serum samples with respect to the degradation of HSA. This means that the levels of degradation product, which in this case of course depend on the concentration of synthetic peptide added to the serum sample, can be higher than the levels of HSA degradation products present in the samples. This then means that detection of the specific fragments is easier and more pronounced. It also means that the ratios between the differentially expressed fragments are less prone to variation than in the case of HSA degradation. Further, it has been found, as can be seen from Fig. 7, that after

Rapigest™ and trypsin treatment of the sample, wherein the Rapigest™ treatment is a known method to the skilled person to add detergents compatible with mass spectrometry, which denature the proteins for a proper digestion by trypsin, the synthetic peptide RHPYFYAPELLFFAK was not degraded in prostate carcinoma patients, while in healthy individuals the C- terminal lysine moiety of said synthetic protein was cleaved off. The ratio between the level of the original synthetic product and the degradation product, as plotted in Fig. 7, is a useful measure to clearly discern between patients and healthy subjects. The synthetic peptide and its degradation product(s) can be easily detected on basis of their molecular mass: the synthetic peptide has a mass of 1898.8 Dalton and the degradation product RHPYFYAPELLFFA has a molecular mass of 1770.9 Dalton.

Thus, another preferred embodiment of the method for detecting a prostate carcinoma with an artificial peptide as discussed above, comprises an additional Rapigest™ and trypsin digestion step in the assay. Also for this step first the optimal conditions should be determined. Generally, about 2 hours incubation for about a 100 fold diluted serum should be sufficient.

A further embodiment of the present invention is checking for effects of treatment on (metastasis of) prostate cancer by regularly performing one of the above described assays on prostate cancer patients who receive treatment. In this way, it can be checked whether the prostate cancer specific proteolytic characteristics of the serum of the patient change into the pattern seen with healthy controls. Such a regular check can also be used to adjust the treatment intensity, regime or pharmaceutical dose.

The invention is now illustrated by the following non-limiting example.

Example Samples

All samples were obtained via the European Randomized study of Screening for Prostate Cancer (ERSPC) biobank, ErasmusMC and are collected under uniform conditions. Frozen serum samples were thawed on ice once, aliquoted and immediately refrozen.

Depletion

Depletion was performed on an immunoaffinity column (Agilent, Santa Clara, USA) according to the recommendations of the manufacturer. Briefly, 25 mL serum was diluted to 125 mL with loading buffer and spin-filtered (0.22 mm) for 20 min at 13 000 rpm in an Eppendorf centrifuge at 4 ⁰C. 75 mL of each sample was loaded onto the column using an autosampler cooled to 4⁰C. Depletion was performed at room temperature on an AKTA FPLC system (GE-He althcare, Chalfont St. Giles, UK) using the following program: 10 min at 100 % eluent A at 0.25 mL/min; 3.5 min at 100 % eluent B at 1 mL/min; 5.5 min at 100 % eluent A at 1 mL/min. Fraction collection was started automatically when a threshold of 25 mAU at 280 nm was exceeded and the collected fractions were snap frozen in liquid nitrogen within 5 min after collection and stored at -80 ⁰C (15).

Serum sample profiling

Pre -fractionation of tryptic digests of the full and depleted serum samples were performed using ClinProt magnetic beads, using hydrophobic interaction MB-HIC 18 (Bruker Daltonik, Leipzig, Germany). Immediately before tryptic digestion, samples were thawed to assure that each sample was treated equally in terms of freeze-and-thaw cycles and incubation/storing time at room temperature. Six μl of each serum sample was diluted 10 times in milliQ to obtain a total volume of 60 μl. To each sample 6 μl of 1 % Rapigest™ (Waters, Milford, USA) dissolved in 50 mM ammoniumbicarbonate was added. Samples were incubated for 2 minutes at 37 ⁰C and subsequently 7.0 μl of 0.1 μg/μl 3 mM Tris HCl gold grade trypsin solution (Promega, Madison, USA) was added and incubated overnight at 37 ⁰C. After overnight incubation, 8 μl of 500 mM HCl was added in order to obtain a final concentration of 50 mM HCl (pH < 2), and incubated further for 45 min at 37 °C. After this preparation procedure samples were used for magnetic bead fractionation.

All magnetic bead preparations were performed according to the manufacturer's instructions. Beads and sample were incubated in binding buffer for several minutes at room temperature. The beads were separated from the supernatant using a magnetic separation device. After washing the bound peptides were eluted by 0.1 % Trifluoroacetic acid/50 % acetonitrile/50 % water.

The eluted peptides were spotted in fourfold onto a MTP AnchorChip 600/384 target plates using α-cyano-4-hydroxycinnamic acid (HCCA) (Bruker Daltonik, Bremen, Germany) as matrix according to the manufacturer's recommendations. MALDI-TOF mass spectra were obtained in the automated AutoXecute mode using an Ultraflex TOF/TOF instrument (Bruker Daltonik, Bremen, Germany), operated in reflectron mode in the mass range of 800-4 000 Da and equipped with a nitrogen laser (wavelength: 366 nm). For all samples, the magnetic bead based sample fractionation and MALDI-TOF target preparation were performed on a fully automated robotic platform (ClinProt Robot, Bruker Daltonik, Leipzig, Germany). This was done to enable high throughput and to assure optimal reproducibility. The whole experiment was performed in threefold with intervals of at least one month allowing a reliable assessment of the reproducibility and to only select differentially expressed peaks that are robust over time (present in 2 out of 3 separate runs).

The magnetic bead fractions of two samples (one control and metastasis) were re-measured by MALDI FTICR MS. One μl of eluted sample was spotted onto an anchorchip target plate (600/384 anchorchip with transponder plate; Bruker Daltonik GmbH, Bremen, Germany). Before the spots were dried, one μl of 2,5-dihydroxy-benzoic acid (DHB) matrix (Bruker Daltonics GmbH, Bremen, Germany), 10 mg/ml in 0.1 % TFA water was added and the spots were allowed to dry at ambient temperature. All spots were measured with an Apex Q 9.4 Tesla MALDI FTICR MS with a combi-source, (Bruker Daltonics, Billerica, USA). For each measurement 100 scans were summed up and for each scan ions generated by 10 laser shots were accumulated.

Reproducibility calculation

The inter- and intra-experimental reproducibilities of the peak intensity were calculated using the ClinProtools software package 2.0 build 365 (Bruker Daltonics, Bremen, Germany). This software tool was used for alignment, normalization and peak selection of each spectrum. A signal to noise of > 2 was used for peak picking, the peak list generated in this way contained for each peak position an average peak intensity and standard deviation. These average intensities and standard deviations were used to calculate the coefficient of variation (CV) for each individual peak position for ten samples (five controls, five metastases). The intra-experimental CV was based on the average of the CV values calculated for all masses present in all four replicate measurements within one experiment. The inter-experimental CV was calculated based on the peak positions present in the twelve measurements of one sample over the three experiments.

Statistical analysis of differences in peptide profiles

The differences in intensities of peptide masses in MALDI-TOF mass spectra of patient and a control group were compared and the differences were statistically analysed with a database application (M.K. Titulaer et al.

(2006). A database application for pre-processing, storage and comparison of mass spectra derived from patients and controls. BMC Bioinformatics 7, 403).

A new version of the database application was used, which was adapted for peak picking of masses above a signal to noise threshold of 4 (D.M. Horn et al. (2000). Automated reduction and interpretation of high resolution electrospray mass spectra of large molecules. J Am Soc Mass Spectrom 11, 320-332). Spectra from the three MS experiments performed with a time interval of 1 month were compared. Each experiment contained 4 replicate spectra of 27 samples of prostate cancer patients and 30 samples of a control group. From the in total 4 replicate spectra of each sample, 3 were used to create a profile matrix of the mean intensity of each peptide mass for each sample (the same exclusion criteria were used as previous described (L.J. Dekker et al. (2005). MALDI-TOF mass spectrometry analysis of cerebrospinal fluid tryptic peptide profiles to diagnose leptomeningeal metastases in patients with breasts cancer. MoI Cell. Proteomics 4, 1341-1349). Each spectrum was internally calibrated on at least 4 masses present of 5 intense albumin masses, 1 296.7045, 1 511.8427, 1 623.7875, 1 639.9377, and 2 045.0953 Da. Spectra with less then 4 of the 5 albumin masses were rejected and not used in the further analyses. Peptide masses present in at least 4 spectra of all samples were clustered within a mass window of 0.5 Da. If a peak mass was not present in a certain spectrum, the background intensity at that point was taken for correct statistical comparison. The Wilcoxon-Mann- Whitney test was performed on each peptide mass in the matrix, comparing the prostate cancer group and control group of experiment 1, the prostate cancer group and control group of experiment 2, the prostate cancer group and control group of experiment 3, respectively. In a second statistical comparison all spectra of experiment 1, 2 and 3 were combined, which gave 12 replicate spectra for each sample. From these replicates a number of 8 spectra where combined to create a profile matrix of the mean intensity of each peptide mass for each sample (which allows a maximum of four rejected spectra per sample if less then 4 spectra were rejected 8 spectra were selected at random). Again peptide masses present in at least 4 spectra of all samples were clustered within a mass window of 0.5 Da. All differentially expressed peaks with a p-value < 0.01 (Wilcoxon-Mann- Whitney test) in at least 2 of the 3 experiments and present with a p-value < 0.01 in the combined experiment were used as candidates for identification.

MALDI FTICR MS data analyses All MALDI-FTICR MS spectra were analysed by Data Analyses

(Bruker Daltonics, USA version 3.4 build 169) and each spectrum was internally calibrated with respect to the most intense albumin peaks in the sample (960.5631, 1 000.6043, 1 149.6156, 1 511.8427, 2 045.0953 m/z). Subsequently, peak picking is performed with the SNAP 2 algorithm with a signal to noise threshold of 4 and a quality factor of 0.9.

Identification by NanoLC MALDI-TOF/TOF

Fractionation was performed using a Monolithic column (200 mm i.d. Dionex, Sunnyvale, CA, USA) on a nanoscale liquid chromatography system (nanoLC) (Dionex, Sunnyvale, CA, USA). The samples in which the differentially expressed peaks had the highest intensity were pooled for identification. From this sample 5 ml of digested samples was loaded onto a trap column (250 mm i.d. x 5 mm, Dionex, Sunnyvale, CA, USA). Fractionation was performed using a 50 minute gradient from 0 % to 64 % of acetonitrile, (solution A (100 % H₂O, 0.05 % TFA) and solution B (80 % ACN, 20 % H₂O and 0.04 % TFA); 0 to 3 min, 0 % solution B, 35 min 45 %, 35.1 min 80 %, 38 min 80 %, 38-50 min 0 % with a flow of 2 μl/min. 10 second fractions were spotted automatically onto a commercially available pre-spotted MALDI plate containing 384 spots (Bruker Daltonics, USA) of α-cyano-4-hydroxycinnamic acid (HCCA) matrix, using a robotic system (Probot Micro Fraction Collector, Dionex, Sunnyvale, CA, USA). To each fraction, 0.33 μl H₂O was added. Finally, we used a 10 mM (NH)₄H₂PO₄ in 0.1 % TFA/H₂O solution to wash the pre-spotted plate for 5 s to remove salts. The plate was subsequently measured by automated MALDI-TOF/TOF (Ultraflex, Bruker Daltonics, Germany) using WARP-LC software (Bruker Daltonics, Germany) which obtained MS spectra of each individual spot and subsequently performed MS/MS. The spots for performing MS/MS measurements were determined automatically by the WARLP-LC software based on signal intensity and presence of interfering peaks. The MS/MS data resulting from the WARP-LC measurements were searched against the Swiss Prot database using the Mascot search engine with a 100 ppm mass tolerance for the parent ion and a 0.6 Da mass tolerance for fragments. All identified differentially expressed peptide peaks were confirmed by comparing the accurate mass of the MALDI-FTMS measurement with the calculated mass for the identified sequence within a mass window of 2 ppm. For some peaks it was not necessary to perform an LC separation and these peaks have been identified by direct MS/MS measurements using the same equipment as described above.

Proteolytic assays using synthetic peptides Synthetic peptides with the sequences listed in Table 1 and having a purity of > 98% were purchased from Pepscan, Lelystad, The Netherlands.

Table 1. Synthetic peptides purchased at > 98 % purity

Control serum samples were used, undiluted, 10, 100 and 1 000 fold diluted, respectively. 1 μl of the diluted serum sample was incubated with 10 μl of 10 pmol/μl synthetic peptide in H2O for 30 min at 37 ⁰C and subsequently 0.5 μl was spotted with 0.5 μl of DHB on an anchorchip plate and measured by MALDI-FTMS. In addition, a time series was performed: 0.5 min, 1 min, 2 min, 5 min, 10 min, 20 min, 40 min, 80 min, 120 min and overnight incubation at 37 ⁰C, respectively. For these experiments a 100 fold diluted serum was used to obtain a similar serum concentration as in the profiling experiment. These samples were also spotted and measured by MALDI-FTMS. Also, negative controls were included in which the peptides were replaced by water or in which the serum was replaced by water. Control samples (n=5) and prostate cancer with metastasis samples (n=5) were diluted 100 fold and subsequently 1 μl of the diluted serum was added to 10 μl of 10 pmol/μl of the synthetic peptide (RHPYFYAPELLFFAK) followed by incubation for 2, 4 and 15 h. The resulting mixtures were measured by MALDI-FTMS. In an additional experiment 50 μl of the 100 fold diluted serum samples were first incubated for two hours with 50 μl of 0.1 % Rapigest™ (Waters, USA) dissolved in 50 mM ammoniumbicarbonate and 10 μl of 0.1 μg/μl 3 mM Tris HCl gold grade trypsin solution (Promega, USA). Subsequently, 1 μl of this solution was added to 10 μl of 10 pmol/μl of the synthetic peptide RHPYFYAPELLFFAK and measured by MALDI-FTMS. This experiment is performed independent in 3 fold. In addition, similar experiments were performed in which the incubation time of the serum with trypsin was increased to 4 h. The linear relation between the amount of sample and the intensities in MALDI-FTMS measurements have been tested by measuring a dilution series of peptide RHPDYSWLLLR (1 467.8430 Da) in the concentrations of 0.5, 1, 2, 5 and 10 pmol/μl spiked with 2 pmol/μl of peptide RHPYFYAPELLFFAK (1 898.9951 Da) in the presence of a 100 fold diluted serum sample. MALDI-FTMS measurements have been performed as described above. Results

A set of 57 tryptically digested serum samples (control n=30, prostate carcinoma with metastasis n=27) were used for a tryptic peptide profiling experiment and the resulting spectra were analy2ed and statistically compared. The reproducibility of peak intensity, intra-experimental and inter- experimental, were calculated, and the average CVs for the peptides present in all spectra were 30 % and 45 %, respectively. In total 2 410 possible peak positions were detected in the analyses of the three combined experiments. Of these 2 410 peak positions 94 were significantly differentially expressed with p < 0.01 in the Wilcoxon-Mann- Whitney test. In Figure 1, the number of peaks is presented for each p-value interval. On the same data, we performed a cross validation by randomly assigning a group number to each serum sample and then repeating the Wilcoxon-Mann- Whitney test. This scrambling procedure was repeated 1 000 times. The average frequency of possible background peaks per p-value interval is presented by the red line in Figure 1. The relative flat distribution of the p-value histogram indicates that there is no correlation between peak positions and groups after scrambling. The p-value histogram of the actual experiment is clearly skewed to lower p-values and is significantly different from the histogram after scrambling (Figure 1). All experiments were also analysed separately and all significant peaks from these analyses were combined; peaks that were present in at least 2 experiments were extracted from the list. It was found that 22 differential expressed peak positions were present in at least two experiments. If present in all 3 experiments this number drops to 18. The 22 peak positions corresponded to 8 mono-isotopic masses. For mass 1 249.6 Da a gel-view presentation is shown in Figure 2 in which a clear difference in intensity between the control and cancer samples can be observed. These 8 mono-isotopic peaks were identified with MALDI-TOF/TOF and nanoLC MALDI-TOF/TOF experiments and identifications were confirmed by performing exact mass measurements of the parent ion with MALDI-FTMS (Table 1) (L.J. Dekker et al. (2006). FTMS and TOF/TOF mass spectrometry in concert: Identifying peptides with high reliability using matrix prespotted MALDI target plates. J Chromatogr B Analyt Technol Biomed Life Sci 847, 62-64). This resulted in the identification of 8 peptides from HSA, 2 of them being tryptic and 6 semi-tryptic. The semi-tryptic peptides are derived from the 2 tryptic peptides (Figure 3). We also checked the MALDI-FTMS mass spectra for additional HSA fragments also derived from these two tryptic peptides; if masses matched within 2 ppm they were included in Figure 3. Also for most of these peptides MS/MS data were obtained as indicated in Table 2.

All fragments in Figure 3 have significant p-values (p < 0.01) in the Wilcoxon-Mann- Whitney test of the combined data. The two tryptic fragments show an overlap in sequence as can be seen from Figure 3c. The average peak intensity and standard deviations of the identified peptides are displayed in Table 2. In Figure 4 the patient/control ratios for the average intensities of the identified peptides are plotted. In this figure one can observe that the intensities of the tryptic fragments are higher in prostate carcinoma patients with metastasis compared to the controls. Conversely, the semi-tryptic fragments are significantly higher in the control samples. In the spectra acquired from both fractions of the depleted samples these specific albumin fragments were not present and the identified peptides were also not observed as naturally occurring peptides in full serum.

Synthetic peptides were purchased with the same sequence as the two HSA peptides that show degradation (RHPDYSWLLLR and RHPYFYAPELLFFAK). In addition, two peptide sequences with an extended c-terminus, an extended n-terminus of the above mentioned peptides were purchased. For background assessment peptides with the reverse sequence were also purchased, see Table 1. In an initial experiment 10 pmol of each peptide is incubated with a 100 times diluted control serum sample for two hours. For both peptides degradation is observed 13 % and 54.5 %, respectively of the original peptide. The observed degradation pattern is similar to that of the profiling experiment, see Figure 5. In the negative control experiment no proteolytic degradation of the synthetic peptide was observed. For the reverse sequence of the peptide only a small amount of peptide is degraded < 5 % (which can not be determined more precisely because of impurities in the synthetic peptides with the same exact mass as the expected fragments). For the peptides with the extended c-terminus, a similar low amount of degradation is observed. Peptides with the extended n-terminus show a clear increase in degradation to 23 % and 93 %, respectively. For the peptide with sequence RHPYFYAPELLFFAK a time series and serum dilution series were performed. With an increasing serum concentration; a higher percentage of degradation is observed, for a 1000, 100, 10 and 0 times dilution of the serum the percentage of the peptide that is degradation is 1.5, 19.2, 92.2 and 95 %, respectively, after half an hour incubation at 37 ⁰C. For the time series a similar effect is observed i.e. more degradation at longer incubation times, see Figure 6. Between control serum samples (n=5) and serum samples of patients with prostate cancer and metastasis (n=5) no significant difference in degradation of RHPYFYAPELLFFAK is observed (p=0.775 Wilcoxon-Mann- Whitney) when 100 fold diluted serum is incubated for 2, 4 and 15 h without the addition of trypsin. In an additional experiment we first incubated the 100 fold diluted serum samples with Rapigest™ and trypsin for 2 h and subsequently added the synthetic peptide followed by incubation for another 2 h. A significant difference in ratio between the synthetic peptide and the degradation products was observed (p=0.005 Wilcoxon-Mann- Whitney) between prostate cancer and control samples Figure 7. The experiment was performed in 3 fold resulting in similar differences in ratio for control serum samples and prostate cancer metastases serum samples. If the serum samples were incubated for 4 h prior to the addition of the synthetic peptide no degradation was observed anymore. Different durations of the incubation of the synthetic peptides with serum trypsin solution were tested, namely 2, 4 and 15 h; for 2 and 4 h similar results were obtained but after an incubation period of 15 h the difference between controls and prostate cancer samples was not longer significant. Note that the degradation products are not present in the purchased synthetic peptide nor are they formed by trypsin. The measurements of a dilution series of peptide RHPDYSWLLLR (1 467.8430 Da) spiked with RHPYFYAPELLFFAK (1 898.9951 Da) in the presence of serum resulted in a linear correlation between the concentration and the relative intensity of peptide RHPDYSWLLLR (R²=0.993).

Figure 1 shows a histogram of p-values where the height of each bar denotes the number of peptide peaks while the horizontal base corresponds to the p-value interval (interval size 0.01) the part of the bar in white represent the number of peaks with a higher intensity in the control and the part in black the peaks with a higher intensity in the prostate cancer with metastasis. The red line represents the histogram of p-values after cross validation. The height of the red line shows the average number of peptide peaks after 1 000 scrambling procedures. The distribution is clearly different from the random distribution (red line) and skewed to the left indicating a high number of peptides that discriminate between the two groups (low p-value) for explanation see Dekker et al. 2005 supra; Titulaer et al. 2006, supra.

This graph shows the distribution of the frequency of peptides as function of their p-value (univariate analyses). If p is small and is elevated above the background line significant differentially expressed it indicates peptides.

Figure 2 a zoom in of mass 1249 Da in a gel view representation of all measured spectra.

Figure 3 shows in panel a and b 2 amino acid sequences that are part of the HSA sequence. The arrows indicate cleavage sites for albumin and the percentage above the arrow the specificity of cleavages. All found and identified peptides that derive from those sequences are shown with their calculated masses (MH). In panel c the homology between the 2 amino acid sequences is shown.

Figure 4 shows on the x-axis the ratio (control/patients) of peak intensities on a logarithmic scale. On the y-axis the sequence of the peptide is indicated.

Figure 5 shows a MALDI-FTMS measurement of a synthethic peptide RHPYFYAPELLFFAK, which is incubated overnight at 37 ⁰C with 100 fold diluted serum. The fragmentation pattern is similar to the fragmentation observed in the profiling experiment.

Figure 6 shows the time dependent proteolytic degradation of synthethic peptide RHPYFYAPELLFFAK. The peptide is incubated with 100 fold diluted serum at 37 ⁰C and subsequently measured by MALDI-FTMS at different time points. In time a clear decrease of mass 1 898 Da (RHPYFYAPELLFFAK) and a clear increase of mass 1 770 Da (RHPYFYAPELLFFA) one of the resulting fragments can be observed in the mass spectra.

Figure 7 shows the differential degradation of synthetic peptide observed between control serum samples and prostate cancer serum samples. The confidence interval of the ratio between the synthetic peptide with mass 1 898.9 Da and the proteolytic product with mass 1 770.9 for 5 tryptic digests of control serum samples and 5 tryptic digest of prostate cancer serum samples incubated with synthetic peptide (RHPYFYAPELLFFAK) for 2 h is shown.

SEQUENCE LISTS

NUMBER OF SEQ ID Nos : 20

SEQ ID NO. 1 LENGTH: 9 TYPE: Peptide

SEQUENCE: 1

His Pro Asp Tyr Ser VaI VaI Leu Leu 1 5

SEQ ID No. 2 LENGTH : 9 TYPE: Peptide

SEQUENCE : 2

His Pro Tyr Phe Tyr Ala Pro GIu Leu 1 5

SEQ ID No. 3 LENGTH: 26

TYPE: Peptide

SEQUENCE : 3

Leu Tyr GIu lie Ala Arg Arg His Pro Tyr Phe Tyr Ala Pro GIu 1 5 10 15

Leu Leu Phe Phe Ala Lys Arg Tyr Lys Ala Ala

20 25

SEQ ID No. 4 LENGTH: 21 TYPE: Peptide SEQUENCE : 4

Tyr GIu Tyr Ala Arg Arg His Pro Asp Tyr Ser VaI VaI Leu Leu 1 5 10 15

Leu Arg Leu Ala Lys Thr

20

SEQ ID No. 5 LENGTH: 11

TYPE: Peptide

SEQUENCE: 5

His Pro Asp Tyr Ser VaI VaI Leu Leu Leu Arg 1 5 10

SEQ ID No. 6 LENGTH: 10 TYPE: Peptide

SEQUENCE : 6

His Pro Asp Tyr Ser VaI VaI Leu Leu Leu 1 5 10

SEQ ID No. 7 LENGTH: 12 TYPE: Peptide

SEQUENCE : 7

Arg His Pro Asp Tyr Ser VaI VaI Leu Leu Leu Arg

1 5 10

SEQ ID No. 8 LENGTH: 11 TYPE: Peptide

SEQUENCE : 8

Arg His Pro Asp Tyr Ser VaI VaI Leu Leu Leu 1 5 10

SEQ ID No. 9 LENGTH: 10 TYPE: Peptide

SEQUENCE : 9

Arg His Pro Asp Tyr Ser VaI VaI Leu Leu 1 5 10

SEQ ID No. 10 LENGTH: 15 TYPE: Peptide

SEQUENCE: 10

Arg His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe Ala Lys 1 5 10 15

SEQ ID No. 11 LENGTH: 10 TYPE: Peptide

SEQUENCE: 11

His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu 1 5 10

SEQ ID No. 12 LENGTH: 11 TYPE: Peptide SEQUENCE: 12

His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe 1 5 10

SEQ ID No. 13 LENGTH : 12 TYPE: Peptide

SEQUENCE: 13

Arg His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe

1 5 10

SEQ ID No. 14 LENGTH: 11 TYPE: Peptide

SEQUENCE : 14

Arg His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu 1 5 10

SEQ ID No. 15 LENGTH: 14 TYPE: Peptide

SEQUENCE: 15

His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe Ala Lys 1 5 10

SEQ ID No. 16 LENGTH: 16

TYPE: Peptide SEQUENCE: 16

Arg His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe Ala Lys 1 5 10 15 Arg

SEQ ID No. 17 LENGTH : 14 TYPE: Peptide

SEQUENCE: 17

Arg His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe Ala

1 5 10

SEQ ID No. 18 LENGTH : 13 TYPE: Peptide

SEQUENCE: 18

His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe Ala 1 5 10

SEQ ID No. 19 LENGTH: 12 TYPE: Peptide

SEQUENCE: 19

His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe 1 5 10

SEQ ID No. 20 LENGTH: 10

TYPE: Peptide SEQUENCE: 20

Arg His Pro Tyr Phe Tyr Ala Pro GIu Leu 1 5 10

SEQ ID No. 21 LENGTH: 15 TYPE: Peptide

SEQUENCE: 21

His Pro Tyr Phe Tyr Ala Pro GIu Leu Leu Phe Phe Ala Lys Arg 1 5 10 15

Table 2: Results of statistical analyses

S

Table 3: Information of identifications and peptides

OJ

Claims

1. Peptide marker for the detection of prostate carcinoma, said peptide marker being a proteolytic degradation product of a serum, wherein said peptide marker comprises an amino acid sequence derived from an amino acid sequence defined by any of the SEQ ID No. 7, 10, 15, 16 and 21.

2. Peptide marker according to claim 1, wherein said peptide marker is a tryptic fragment of the human serum albumin (HSA) protein.

3. Peptide marker according to any one of the preceding claims, wherein said peptide marker comprises any one of the amino acid sequences defined by SEQ ID No. 5-21.

4. Peptide marker according to any one of the preceding claims, wherein the peptide marker is a fragment of an amino acid sequence of HSA defined by SEQ ID No. 7, 10, 15, 16 or 21.

5. Peptide marker according to claim 4, wherein said serum is obtained from a prostate cancer patient with metastasis.

6. Method for detecting a prostate carcinoma comprising obtaining a first serum sample from a subject potentially suffering from prostate carcinoma and optionally a second serum sample from a subject not suffering from prostate carcinoma; subjecting said first serum sample and said optional second serum sample to tryptic digestion to obtain a first tryptic digest and an optional second tryptic digest, respectively; analysing said first tryptic digest and said optional second tryptic digest for expressed peptides by peptide profiling; examining said first tryptic digest and said optional second tryptic digest for one or more differentially expressed peptides; and optionally comparing the results obtained with a control or panel of control samples; and assign a diagnosis on basis of the levels of the peptide fragments.

7. Method according to claim 6, wherein the analysing said tryptic digests is preceded by purification of the tryptic digests with surface active magnetic beads.

8. Method according to claim 6 or 7, wherein analysing said tryptic digests comprises identification of the expressed peptides by MALDI-TOF, nanoLC-MALDI-TOF/TOF and/or MALDI-FTMS.

9. Method according to any one of claims 6-8, wherein said one or more differentially expressed peptides comprise a peptide marker according to any one of claims 1-5.

10. Method for detecting a prostate carcinoma comprising obtaining a serum sample from a subject potentially suffering from prostate carcinoma and optionally a control sample; measuring the specific protease activity by adding a peptide having the sequence of SEQ ID No:7 or 10 analysing the peptide content of the sample after the protease reaction - assign a diagnosis on basis of the levels of the peptide fragments..

11. Method according to claim 10, wherein additionally said serum sample is pretreated with trypsin and Rapigest™.

12. Method for determining the efficacy of a prostate carcinoma treatment comprising periodically performing any of the methods according to claim 6-12 for prostate carcinoma patients who receive treatment.