SUGAR CHAIN SPECIFIC TO PROSTATE CANCER, AND TEST METHOD USING SAME

Abstract

Provided is a test method for identifying prostate cancer by analyzing a sugar chain modifying PSA in a specimen, and detecting an abundance of a multisialylated LacdiNAc structure, in particular Glycan ID: 7512, Glycan ID: 7603, Glycan ID: 7612, and/or Glycan ID: 7613. Furthermore, calculation of PSA G-index from relative abundance(s) of Glycan ID: 7512 and/or Glycan ID: 7603 enables detection of prostate cancer with good specificity even in a patient having a PSA value in a gray zone.

Claims

1. A test method of prostate cancer, comprising: analyzing a sugar chain modifying PSA in a specimen; and analyzing a multisialylated LacdiNAc structure.

2. The test method of prostate cancer according to claim 1, comprising: analyzing a sugar chain modifying PSA in a specimen; and analyzing relative abundance(s) of Glycan ID: 7512, Glycan ID: 7603, Glycan ID: 7612, and/or Glycan ID: 7613.

3. The test method of prostate cancer according to claim 2, wherein the relative abundances of Glycan ID: 7512 and Glycan ID: 7603 are analyzed by logistic analysis.

4. The test method of prostate cancer according to claim 2, comprising: substituting the relative abundances of Glycan ID: 7512 and Glycan ID: 7603 into the following formula 1:
[Expression 1]
log.sub.e(p/(1−p)=−5.85+4.72x.sub.1+0.80x.sub.2   (Formula 1) wherein p represents a value of PSA G-index; x.sub.1 represents a relative abundance of Glycan ID: 7512, and x.sub.2 represents a relative abundance of Glycan ID: 7603.

5. The test method of prostate cancer according to claim 1, wherein the specimen is blood, serum, plasma, or urine.

6. The test method of prostate cancer according to claim 1, wherein the specimen is a specimen obtained from a patient having a PSA value of 4 to 10 ng/ml.

7. The test method of prostate cancer according to claim 1, wherein the sugar chain is analyzed using an oxonium monitoring method.

8. (canceled)

9. A method for testing a grade of prostate cancer, comprising: analyzing a sugar chain modifying PSA in a specimen; and analyzing relative abundance(s) of at least one or more of Glycan ID: 7512, Glycan ID: 7603, Glycan ID: 3401, and/or Glycan ID: 5602.

10. The test method according to claim 9, wherein the specimen is blood, serum, plasma, or urine.

11. The test method according to claim 9, wherein the analysis uses an oxonium monitoring method.

12. The test method of prostate cancer according to claim 1, wherein the analysis uses an antibody or lectin.

13. The test method of prostate cancer according to claim 12, wherein the method uses an antibody or lectin that specifically recognizes a saccharide having a multisialylated LacdiNAc structure.

14. A test kit of prostate cancer, comprising an antibody or lectin that specifically recognizes a saccharide having a multisialylated LacdiNAc structure.

15. The test kit of prostate cancer according to claim 14, wherein the saccharide having a multisialylated LacdiNAc structure is Glycan ID: 7512, Glycan ID: 7603, Glycan ID: 3401, or Glycan ID: 5602.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0029] FIG. 1 is a diagram showing sugar chain structure analysis of PSA using an oxonium monitoring method. FIG. 1A shows an overview of the oxonium monitoring method. FIGS. 1B and 1C show the analysis results with PSA standard.

[0030] FIG. 2 is a diagram showing sugar chain analysis of PSA clinical samples. FIG. 2A shows types and abundance of sugar chains detected in a prostatic hypertrophy patient group and a prostate cancer patient group of a training set. FIGS. 2B to 2D show relative abundances of sugar chains having a LacdiNAc structure detected in the prostatic hypertrophy patient group and the prostate cancer patient group. FIGS. 2E and 2F show the results of Erexim® analysis of representative sugar chains modifying PSA in serum of prostate cancer patients.

[0031] FIG. 3 is a diagram showing the results with PSA G-index. FIG. 3A shows relative abundances of Glycan ID: 7512 in a prostatic hypertrophy patient group and a prostate cancer patient group and FIG. 3B shows relative abundances of Glycan ID: 7603 in the prostatic hypertrophy patient group and the prostate cancer patient group. FIG. 3C shows a plot of PSA G-index values for the prostatic hypertrophy patient group and the prostate cancer patient group of a training set. FIG. 3D shows a plot of PSA G-index values for the prostatic hypertrophy patient group and the prostate cancer patient group of a validation set. FIG. 3E shows ROC curves with PSA, PSA f/T, and PSA G-index values. It should be noted that Glycan ID defines the type of glycan, and the digits correspond to the numbers of HexNAc, Hexose, Fucose, and Neu5Ac from the left, respectively, which are described later in detail.

[0032] FIG. 4 is a diagram showing sugar chains capable of classifying the grade (malignancy) of prostate cancer. FIGS. 4A, 4B, 4C, and 4D show correlations of Glycan ID: 7512, Glycan ID: 7603, Glycan ID: 3401, and Glycan ID: 5602, with a Gleason score, respectively.

[0033] FIG. 5 is a diagram showing the analysis results by lectin tissue staining. FIG. 5A shows prostate tissue staining results with HE, WFA, and MAM in prostatic hypertrophy patients and prostate cancer patients. FIG. 5B shows prostate tissue staining intensity with WFA in prostatic hypertrophy patients and prostate cancer patients. FIG. 5C shows prostate tissue staining intensity with MAM in prostatic hypertrophy patients and prostate cancer patients.

DESCRIPTION OF EMBODIMENTS

[0034] The analysis by the present inventors has revealed that there is a significant increase in certain sugar chains having LacdiNAc (GalNAcβ1-4GlcNAc, N-acetylgalactosamine-N-acetylglucosamine) structure in patients suffering from prostate cancer. The analysis has been performed by mass spectrometry, but any method can be used as long as it is possible to recognize or analyze the multisialylated LacdiNAc structure or the specific sugar chain indicated by Glycan ID below.

[0035] The analytical method by mass spectrometry is a method that can be examined with great sensitivity as shown below, but not limiting to mass spectrometry, the analysis can be performed using anything that recognizes a particular sugar chain, such as an antibody, lectin, or aptamer which recognizes a multisialylated LacdiNAc structure. When the detection is performed with an antibody, the detection may be performed by any method used in the art, such as ELISA or SPR. When the detection is performed with a lectin, the detection may be performed by any method such as lectin blot or capillary electrophoresis. As shown in the Examples below, lectins such as WFA or MAM does not specifically recognize a multisialylated LacdiNAc structure. However, a system capable of specifically detecting a multisialylated LacdiNAc structure can be created by using several lectins in combination.

[0036] The present invention can be also carried out not only by analysis of relative abundances of Glycan ID: 7512 and Glycan ID: 7603, but also by profile recognition on profiling data of saccharides containing a multisialylated LacdiNAc structure. Specifically, profiles of multiple sugar chains obtained from patients having diagnosis determined prostate cancer are machine-learned by a computer as training data, and used it as a diagnostic support system. Then, during the test, by inputting the profile obtained from the subject as it is, prostate cancer candidate data may be detected by pattern recognition.

[0037] As shown in the Examples below, as a result of analyzing sugar chain structures modifying PSA, it has been revealed that prostate cancer can be detected with high accuracy by using PSA G-index defined below. Examination with PSA G-index on patients of a gray zone in conventional PSA tests as a secondary screening makes it possible to detect patients suffering from prostate cancer in a non-invasive manner. Further details are described below with showing data.

[0038] First, whether sugar chains modifying PSA are able to be sensitively detected and quantified was confirmed with PSA standard. PSA obtained from human semen was dissolved in 8 M urea, 50 mM HEPES-NaOH, pH 8.0, reduced with 10 mM DTT (GE Healthcare Life Science), alkylated with 25 mM iodoacetamide (Sigma-Aldrich), and then digested with Trypsin/Lys-C mix (Promega Corporation). Glycoproteins were enriched by hydrophilic purification method (Non Patent Literature 7), and dried in vacuo.

[0039] The resulting glycoproteins were analyzed by oxonium ion monitoring method (Erexim method, Non Patent Literature 8, Patent Literature 9) with a triple quadrupole mass spectrometer LCMS-8060 (Shimadzu Corporation) coupled with a Prominence nanoflow liquid chromatogram to identify sugar chains.

[0040] FIG. 1A schematically shows an overview of the oxonium ion monitoring method applied by multiple-reaction monitoring mass spectrometry (MRM-MS). PSA is degraded with trypsin into a mixture of peptides and glycopeptides, and they are separated with a nanoflow liquid chromatogram. In a mass spectrometer, glycopeptide ions having a particular mass are isolated at the first quadrupole (Q1). The isolated glycopeptide ions are introduced into the second quadrupole (Q2), and cause collision induced dissociation (CID). In the third quadrupole (Q3), oxonium ions from saccharides are selectively detected with a filter. Under optimal collision energy conditions, oxonium ions of m/z 138.1 function as quantitative reporters of glycopeptides of various saccharide compositions.

[0041] The results of performing sugar chain structure analysis with 100 ng of PSA standard obtained from human semen are shown in FIG. 1B. The digits on the horizontal axis in the figure define the type of each glycan as Glycan ID, and correspond to the numbers of HexNAc, Hexose, Fucose, and Neu5Ac from the left, respectively. The type and quantitative distribution of saccharides modifying asparagine at position 45 of PSA were analyzed. As a result, it was revealed by oxonium ion monitoring method that there were 67 types of saccharide structures.

[0042] Oxonium ion monitoring method has been found to be not only a method with which the analysis was completed in a short time of 25 minutes of LS/MS operation time without the need for enzymatic separation of glycans or chemical modification, but also a very sensitive detection method. The saccharide of the highest content was 4[HexNAc]5[Hex]1[Fuc]2[Neu5Ac] (Glycan ID: 4512, 44.5±0.9%), and the saccharide of the lowest content was 6[Hex]7[NAcHex]0[Fuc]0[Neu5Ac] (Glycan ID: 6700, 0.01±0.001%). There has been previously no report of detecting such a very large number of types of saccharides as 67 types of glycans modifying PSA, or detecting a very small amount of saccharide as 0.01%. Thus, it has been shown that oxonium ion monitoring method is a very sensitive method.

[0043] FIG. 1C shows the dynamic range of PSA sugar chain analysis. PSA digested with trypsin was serially diluted, and the detection limit was analyzed. As a result, it was possible to detect up to 1 fmol of the glycopeptide obtained from 0.03 ng of PSA, and the quantitative dynamic range was 5 digits or more (R.sup.2>0.99). This result indicates that sugar chain analysis by oxonium ion monitoring method can perform test sufficiently even with PSA in a gray zone of 4 to 10 ng/ml.

[0044] Next, sugar chain analysis of PSA was performed using samples obtained from 15 prostate cancer patients and 15 prostatic hypertrophy patients having a PSA value in a gray zone, 4 to 10 ng/ml, to determine an index identifying prostate cancer, as a training set. Table 1 shows patient characteristics in the training set.

[0045] Analysis with clinical samples uses serum samples of prostate cancer patients and prostatic hypertrophy patients. It should be noted that final diagnosis is determined from histopathological diagnosis by prostate biopsy. Purification of PSA was performed as follows. To the serum was added 4-fold amount of wash solution (0.1% Tween-20 in PBS). The obtained solution was mixed with Protein G Sepharose to which anti-PSA monoclonal antibodies were immobilized (Fitzgerald), reacted overnight at 4° C., and washed. Then, PSA was eluted with 8 M urea, 50 mM HEPES-NaOH, pH 8.0, and purified. The eluted protein was reduced with 10 mM DTT, alkylated with 25 mM iodoacetamide, and then digested with Trypsin/Lys-C mix. Glycoproteins were enriched by hydrophilic purification method (Non Patent Literature 7), and dried in vacuo. It should be noted that, although serum samples were used here, not only blood derived samples but also urine can be used as samples.

TABLE-US-00001 TABLE 1 Training set Validation sample set Prostate Prostatic Prostate Prostatic Characteristics cancer hypertrophy cancer hypertrophy Number of 15 15 15 15 patients Age, 69.1 ± 6.69 68.1 ± 5.0 70.3 ± 3.3 69.9 ± 6.4 Mean ± SD (53-79) (60-79) (65-75) (58-79) (range) PSA (ng/mL) 7.5 ± 1.6 7.4 ± 1.5 7.5 ± 1.5 7.5 ± 1.4 Mean ± SD (4.19-9.76) (4.31-9.36) (5.41-9.67) (5.32-9.43) (range) f/T PSA (%) 16.0 ± 6.4 20.8 ± 7.7 16.1 ± 5.9 19.3 ± 7.5 Mean ± SD (8.4-34.2) (9.8-45.1) (9.2-27.8) (9.8-34.0) (range)

[0046] In the same manner as described above, 52 sugar chain structures on PSA could be quantified (FIG. 2A). Sugar chain structures of Glycan ID: 7512, 7603, 7612, and 7613 showed significant quantitative differences between the prostatic hypertrophy patient group and the prostate cancer patient group. It has also been found that sugar chains modifying PSA differ between the prostate cancer patient group and the healthy subject group, although no data is shown here. Thus, it is possible to distinguish prostate cancer patients from others, namely patients suffering from prostate diseases other than prostate cancer and healthy subjects, by analyzing sugar chain structures.

[0047] Among the 52 types of sugar chains, sugar chains containing a multisialylated structure, specifically di-/tri-sialylated LacdiNAc (GalNAcβ1-4GlcNAc), were significantly increased in the prostate cancer patient group compared to the prostatic hypertrophy patient group (FIG. 2 (D), p=0.0023). Meanwhile, the total amount of sialylated LacdiNAc (FIG. 2B) or sugar chain structures to which mono-sialylated LacdiNAc is attached (FIG. 2C) did not show significant changes between the prostatic hypertrophy patient group and the prostate cancer patient group. The terminal LacdiNAc structure and the presence of sialic acid (Neu5Ac) residues were also confirmed by Erexim method (FIGS. 2E and 2F).

[0048] Based on these findings, we aimed to establish a novel diagnostic algorithm that complements the specificity of PSA test and can reliably reduce false positive rates. Two sugar chain structures specific for prostate cancer, Glycan ID: 7512 (p=9.91×10.sup.−8) and Glycan ID: 7603 (p=1.66×10.sup.−5), which showed significant differences between the prostate cancer patient group and the prostatic hypertrophy patient group in the training set, were selected, and the relative abundance thereof were plotted. FIG. 3A shows the relative abundances of Glycan ID: 7512 in prostatic hypertrophy patients (BPH) and prostate cancer patients, and FIG. 3B shows the relative abundances of Glycan ID: 7603 in those. Both Glycan ID: 7512 and 7603 show significant difference in abundance between them.

[0049] Based on these results, a diagnostic model (PSA G-index) based on logistic regression was established. In formula 1, p represents a value of PSA G-index; x.sub.1 represents a relative abundance of Glycan ID: 7512; and x.sub.2 represents a relative abundance of Glycan ID: 7603.

[Expression 3]

log.sub.e(p/(1−p)=−5.85+4.72x.sub.1 0.80x.sub.2   (Formula 1)

[0050] When a cutoff value of PSA G-index was set to 0.5, the sensitivity and specificity of the training set were 93.3% and 100%, respectively (FIG. 3C). It should be noted that, since PSA G-index is based on logistic analysis, the formula may differ slightly depending on the increase in the number of specimens to be analyzed in the future. However, it is possible to identify prostate cancer and other prostate diseases with high sensitivity and specificity by detecting Glycan ID: 7512 and Glycan ID: 7603 and using their amounts for identification.

[0051] Although all the saccharides modifying PSA are analyzed and their relative amounts are determined here, analysis may be performed of only a particular saccharide, such as a saccharide having a multisialylated LacdiNAc structure. In addition, among the peptides of PSA, peptides in the region that is not glycosylated may be taken as a reference, and a ratio of PSA modified by a particular saccharide to total PSA may then be determined to use for analysis. Although the relative values of saccharides described above differ depending on the saccharide used as a reference or the amount of PSA, the amounts of the saccharides are significantly different in prostate cancer and other prostate diseases, and thus it is possible to distinguish prostate cancer and other prostate diseases with high sensitivity by logistic analysis.

[0052] The PSA G-index was then evaluated using the validation sample set of Table 1 (FIG. 3D). All clinical samples were diagnosed as correct disease in both of the 15-case prostatic hypertrophy patient group and the 15-case prostate cancer patient group. In ROC curve analysis, the PSA G-index area under the curve (AUC) was 1.00 (100% sensitivity and 100% specificity), while AUCs of the PSA value (Total PSA) and the ratio of free PSA to total PSA (PSA f/T) were 0.50 (80.0% sensitivity, 33.3% specificity) and 0.60 (73.3% sensitivity, 60.0% specificity), respectively (FIG. 3E). These results indicate that PSA G-index can significantly improve specificity of prostate cancer diagnosis and avoid false positives compared to PSA or PSA f/T values, which are conventionally indexes of prostate cancer.

[0053] Next, whether sugar chain structures on PSA in serum reflect the grade (malignancy) of prostate cancer was analyzed. Using serum from 77 patients of different grades shown in Table 2, sugar chain analysis by oxonium ion monitoring method was performed.

TABLE-US-00002 TABLE 2 Prostatic Characteristics hypertrophy GS6 GS7 GS8 GS9 Number of 30 8 23 11 5 patients (range) Age, 69.1 ± 5.8 73.8 ± 3.4 68.9 ± 5.2 72.5 ± 4.5 75.2 ± 6.2 Mean ± SD (58-79) (67-79) (53-78) (67-78) (65-84) (range) PSA (ng/mL) 7.4 ± 1.4 9.7 ± 4.7 10.7 ± 7.0 14.9 ± 1.4 50.8 ± 33.9 Mean ± SD (4.31-9.43) (5.57-21.14) (5.5-27.85) (5.41-28.67) (6.88-107.54) (range) f/T PSA (%) 20.1 ± 7.6 19.5 ± 8.9 14.2 ± 5.6 14.8 ± 8.3 11.7 ± 6.2 Mean ± SD (9.8-45.1) (6.3-34.2) (7.0-26.7) (9.3-34.9) (7.3-24.1) (range)

[0054] As a result, the frequency of Glycan ID: 7512 (FIG. 4A) or Glycan ID: 7603 (FIG. 4B), which are sugar chain structure to which multiple sialylated LacdiNAc are attached, was positively correlated with Gleason score (GS), an index of malignancy by prostate biopsy (p=3.34×10.sup.−8, or p=2.56×10.sup.−9). The Gleason score is an index that classifies the malignancy of prostate cancer that occurs with different degrees of malignancy in the same prostate. The Gleason score is classified into 9 stages of GS2 to GS10 by the sum of the values obtained by determining predominant and ancillary lesions from biopsy. The higher the Gleason score value indicates the higher the malignancy.

[0055] Furthermore, Glycan ID: 3401 (FIG. 4C) or Glycan ID: 5602 (FIG. 4D) showed a negative correlation with the Gleason score (p=1.69×10.sup.−6, or p=9.66×10.sup.−7). These results indicate the possibility that the pathological malignancy of prostate cancer is correlated with the amounts of certain sugar chains, and the malignancy of prostate cancer can be diagnosed by liquid biopsy.

[0056] To identify the cause of the increase of multisialylated LacdiNAc structure in serum PSA, lectin histochemical staining was performed using a tissue microarray (US Biomax) containing 9, 31, and 111 prostate tissues of normal, prostatic hypertrophy patients, and prostate cancer patients, respectively. Histochemical staining was performed using WFA, a lectin that detects a LacdiNAc structure, or MAM, a lectin that detects a Siaa2-3Gal structure. Biotinylated lectins were used for both WFA and MAM, and the analysis was performed using a Vectastain Elite ABC kit (Vector Laboratories). As a result, cytoplasmic and protoplasmic membranes of cancer cells were significantly stained (FIG. 5A). The staining intensity was classified into three levels, “high”, “moderate”, and “low”, and the stained images of tissues of prostate cancer patients and those of normal and prostatic hypertrophy patients were compared. In the prostate cancer tissues, images of “high” in staining intensity were observed at a high frequency of 9.00-fold when stained with WFA and 2.24-fold when stained with MAM (FIGS. 5B, 5C).

[0057] WFA and MAM lectins specifically recognize GalNAc structures containing a LacdiNAc structure and a Neu5Acα2-3Gal structure, respectively. The above results indicate that prostate cancer tissue tends to strongly express glycoproteins containing LacdiNAc and Neu5Ac structures. These histological findings were consistent with the results of PSA sugar chain structure analysis by mass spectrometry. Accordingly, it is possible to determine prostate cancer, and even the stage of prostate cancer, by analyzing sugar chains of PSA in blood without performing tissue biopsy.

INDUSTRIAL APPLICABILITY

[0058] Test using the diagnostic marker, PSA G-index, shown in the Examples, performed as a secondary screening of patients diagnosed as having a suspected prostate cancer from the PSA value, can identify prostate cancer with good specificity. The test can identify even prostate cancer at an early stage with good specificity, which makes it possible to lead to early treatment.