Med Oncol (2014) 31:897 DOI 10.1007/s12032-014-0897-4
ORIGINAL PAPER
Evaluation of histidine-rich glycoprotein tissue RNA and serum protein as novel markers for breast cancer Marwa Matboli • Sanaa Eissa • Hebatallah Said
Received: 4 October 2013 / Accepted: 17 February 2014 Ó Springer Science+Business Media New York 2014
Abstract Advances in the field of breast cancer (BC) biomarkers discovery facilitate diagnosis and treatment of BC in its pre-invasive state. While the genetic tissue markers are making significant advances in understanding the molecular basis of BC, serum has long been considered a rich source for biomarkers. So, integrated genomic and proteomic strategies play a huge role in the analytical validation of BC biomarkers and represent a true milestone in the areas of diagnostics and personalized medicine. This study included 60 cases (BC), 30 patients with fibroadenoma and 30 healthy women. Histidine-rich glycoprotein RNA (HRG) tissue expression was analyzed through gene expression-based outcome for breast cancer online algorithm (GOBO) bioinformatic analysis. To confirm our informatics analysis, HRG RNA was detected in breast tissue samples by RT-PCR, and HRG serum protein was estimated by ELISA. GOBO analysis revealed increased HRG RNA expression in all subtypes of BC with relative higher expression in basal subtype and grade 2. We confirmed these findings by HRG tissue RNA with 71.7 % sensitivity and 93.3 % specificity. HRG serum protein was 86.7 % sensitivity and 80 % specificity. HRG tissue RNA and serum protein could be considered as promising novel markers for prediction of BC prognosis.
Electronic supplementary material The online version of this article (doi:10.1007/s12032-014-0897-4) contains supplementary material, which is available to authorized users. M. Matboli S. Eissa (&) H. Said Oncology Diagnostic Unit, Medical Biochemistry and Molecular Biology Department, Faculty of Medicine, Ain Shams University, P.O. Box 11381, Abbassia, Cairo, Egypt e-mail: [email protected]; [email protected]; [email protected]
Keywords Breast carcinoma Bioinformatics Histidinerich glycoprotein ELISA Serum markers Abbreviations HRG Histidine-rich glycoprotein BC Breast cancer BMI Body mass index LN Lymph node ER Estrogen receptor PR Progesterone receptor Her-2 neu Human epidermal growth factor receptor 2 HT Hormonal therapy OCT Oral contraceptive therapy
Introduction Breast cancer is by far the most common cancer among women of both developed and developing countries accounting for 22.9 % of all female cancers. It is also the main cause of cancer death in females responsible for 13.7 % of their cancer-related mortality. In Egypt, breast cancer is estimated to be the most common cancer among females accounting for 37.7 % of their total cancer cases. It is also the major cause of cancer-related mortality accounting for 29.1 % of their cancer-related deaths [1]. The most efficacious screening tool utilized in the clinic is mammography, leading to great improvement in mortality reduction though lesions \0.5 cm in size remains undetectable by such technology [2]. There is also resistance for seeking such services yearly; many false positives are encountered, and cost-to-benefit ratio of these imaging methodologies is high [3]. Breast cancer subtyping has turned out to be the golden era toward new therapeutic strategies. For example, breast
123
897 Page 2 of 7
cancer basal subtype is characterized by high histologic grade, the absence of HER-2 (ErB2) and receptors for estrogen and progesterone with no approved targeted therapies yet exist, since they do not express any known target proteins [4–7]. Histidine-rich glycoprotein was initially identified through former proteomic studies for breast cancer biomarkers identification [8]. HRG is a2-glycoprotein synthesized by liver and present in plasma and platelets. HRG is potentially involved in numerous biologic processes. It can inhibit rosette formation and interacts with heparin, thrombospondin and plasminogen. There have been conflicting reports on both enhancing and inhibitory effects of HRG on angiogenesis [9, 10]. In this study, we explore a two-stage strategy for evaluation of a serum biomarker corresponding to specific cancer-associated gene over-expression and its application in breast cancer.
Med Oncol (2014) 31:897
who underwent plastic breast surgery (mean age 50.47 ± 11.64 years and range from 36 to 77 years) with matching age and sex to the patient’ groups. Study population demographic and clinical characteristics are shown in (Supplementary table 1). Sample processing Sera were collected before surgery following a standardized protocol and stored within 15 min at -80 °C till sample preparation. A smaller volume aliquots are initially prepared to minimize effect of freeze–thaw cycles and provide multiple replicates that were processed in an identical manner. After primary surgery, a representative part of the tumor tissue was immediately frozen stored at -80 °C until RNA extraction. The samples were coded, and clinical information was unavailable to the technicians performing the mRNA quantification. Bioinformatics analysis of HRG
Materials and methods Study population (Supplementary table 1) This pilot study was approved by Ain Shams Faculty of Medicine ethical committee. A prospective analysis was performed on 120 Egyptian subjects, and 90 patients were selected from The General Surgery department, Ain Shams University Hospitals, Egypt, between November 2010 and June 2012, after obtaining informed consent. Patients presented with breast mass and/or bleeding per nipple were included in the study, and patients with past history of breast cancer or any other malignancy or inflammatory mastitis were excluded from the study. All patients provided paired tissue and blood samples. Breast cancer specimens were obtained after modified radical mastectomy for operable primary breast cancer. The pathology specimens were examined independently by histopathologists to grade and subclassify the tumors according to established criteria [11, 12]. Subsequently, the 90 patients included in the study were classified into malignant and benign groups. The malignant group included 60 patients (mean age 52.47 ± 13.2 years and range from 20 to 81 years). Of those patients, 43 were diagnosed by histopathology with invasive duct carcinoma (IDC), 1 patient with invasive lobular carcinoma (ILC), 9 with mixed invasive duct and lobular carcinoma, and 7 patients with rare histological types. Tumor staging and grading were determined according to TNM and World Health Organization classification [13, 14]. The benign group included 30 patients with benign breast diseases: fibroadenoma (49.6 ± 14.11 years and range from 20 to 67 years). The third study group was represented by 30 healthy volunteers
123
Histidine-rich glycoprotein was selected from a large set of signature peptides identified by Schaub et al. [8]. This includes the fragments with the highest significant relation to breast cancer, followed by bioinformatic analysis of HRG transcript expression using data from the gene expression-based outcome for breast cancer online algorithm (GOBO). Sample processing Sera and breast tissues were collected following a standardized protocol (supplementary files). Total RNA was purified with TriFastTM (Pe QLab Biotechnologie GmbH Corporation) according to the manufacturer’s instructions [15]. RT-PCR for HRG RNA RT-PCRs were done in Hybaid thermal cycler [Thermo Electron (formerly Hybaid) Waltham, MA, USA] using Qiagen one-step RT-PCR Kit (Cat. No. 210210) using HRG primers [16–18] (accession NM_000412.2) (more details on supplementary files), yielding an amplicon of 404 bp separated on 2 % agarose gel stained with ethidium bromide. The quality of RNA was checked by amplification of c-DNA for b-actin to ensure successful completion of c-DNA using primer pair [19] (Accession NM_001017992.2), yielding an amplicon of 309 bp (Fig. 2, supplementary figure 3). HRG protein quantitative estimation Sera from all subjects were subjected to immunoassay to measure serum level of HRG protein by using Enzyme-
Med Oncol (2014) 31:897
linked immunosorbent assay kit for HRG supplied by USCN, Life Science Inc., (China). Statistical analysis Univariate analyses were performed using a chi-square test of association of categoric variables. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated according to standard statistical methods. All analyses were performed using Statistical Package for the Social Sciences software (SPSS Inc., Chicago, IL, USA). The threshold value for optimal sensitivity and specificity of HRG was determined by receiver operating characteristics (ROC) curve [20].
Page 3 of 7 897
breast cancer cases (Table 1; Fig. 1 and supplementary figure 3). There was a statistically significant difference in the expression of HRG among the three groups (P = 0.000) (Table 1). No significant correlation was found between HRG mRNA positivity rate and any of the studied clinicopathological factors except molecular subtype and grade (P [ 0.05) (Supplementary table 2). Results demonstrated no association with ER? tumors (55.8 %) and a higher association with basal subtypes (100 %) and grade 2 tumors (68 %). These data are consistent with the correlations recorded in the GOBO data (Supplementary Figure 1a,b). On the other hand, we failed to find HRG mRNA in all investigated serum samples using RT-PCR. Serum HRG protein analysis
Results HRG transcript expression GOBO bioinformatic analysis Bioinformatics analysis of HRG transcript expression was assessed using data from GOBO [21]. GOBO is a web-based analysis tool that utilizes Affymetrix gene expression data collected from 1,881 breast cancer patients with associated stage, grade, nodal status and intrinsic molecular classification based on the paradigm first reported by the Perou Laboratory [22]). Expression of HRG showed statistically significant difference between different subtypes (P \ 0. 001) with higher expression in basal subtypes than other subtypes. Moreover, significant HRG RNA expression in relation to grade (P = 0.00011) with higher expression in grade 2 tumors were compared with grade 1 and 3 (Supplementary figure 1a-c). The lowest HRG transcript expression was observed in HER2-enriched. We performed through GOBO tool a series of Kaplan–Meier analyses to determine whether HRG expression is associated with overall survival (OS). Positive tissue HRG RNA expression is associated with survival when considering all breast tumor subtypes together, and it is highly associated with poor OS in normal like (P = 0.0023), Lymph node– ER? tumors (P = 0.004888) and grade 2 (P = 0.0079) (Supplementary figure 2a-e). Breast tissue HRG mRNA expression Results from the GOBO transcript expression analysis motivated our research group to investigate HRG mRNA at the gene level in addition to the protein level in sera of breast cancer patients. HRG was validated by performing RT-PCR (reverse transcription polymerase chain reaction) for HRG mRNA expression in the breast biopsies. HRG mRNA was not detected in 30 normal breast tissues, but was detected in 13.3 % (4/30) of fibroadenoma cases and 71.7 % (43/60) of
Histidine-rich glycoprotein protein level was significantly higher in the malignant group (Mean rank = 55. 62) as compared to benign (Mean rank = 30.93) and normal control groups (Mean rank = 19.60), P \ 0.001; ROC curve was constructed to determine the best cutoff value of HRG protein that discriminates malignant from non-malignant groups (Fig. 2). The cutoff value was set at 3.1 mg/dl, and the positivity rate of serum HRG protein (No of cases [3.1 mg/dl) was estimated accordingly among all the study groups (Table 1). In the healthy control group, HRG serum protein positivity rate was (13.3 %), in benign group (26.3 %), while the malignant group expressed the highest positivity rate (86.7 %) with a highly statistically significant difference between the three groups (P = 0.000). No significant correlation was found between HRG protein positivity rate and any of the studied clinicopathological factors except grade, menopausal history, and oral contraceptive treatment (OCT) P [ 0.05 (Supplementary table 2, Fig. 3). HRG serum protein in different breast cancer molecular subtypes HRG serum protein was not associated with ER? tumors, but was highly associated with basal subtypes (all of the basal subtypes show over-expression of serum HRG protein 30/30 (Supplementary Figure 1a,b, Supplementary table 2). These data are consistent with the correlation observed in the GOBO data. Interestingly, grade 2 tumors were associated with a higher serum HRG protein positivity (Supplementary Figure 1c). Sensitivity, specificity, positive predictive value, negative predictive value and accuracy of investigated parameters Sensitivity, specificity, PPV, NPV and accuracy of HRG tissue RNA, and serum HRG protein were tested for their
123
897 Page 4 of 7 Table 1 HRG tissue RNA and serum protein level and positivity rate in the malignant group compared with benign and normal control groups
Med Oncol (2014) 31:897
Group
Serum HRG level
No. of HRG protein [3.1328 mg/dl (%)
No. of HRG RNA positive in breast tissue (%)
Median
1.1333 mg/dl
4/30 (13.3 %)
0/30 (0 %)
Range
0.7107-52.9000 mg/dl
Mean rank
19.6 8/30 (26.7 %)
4/30 (13.3 %)
52/60 (86.7 %)
43/60 (71.7 %)
Normal control (n = 30)
Benign (n = 30) Median
1.5600 mg/dl
Range Mean rank
0.7107–52.9000 mg/dl 30.93
Malignant (n = 60) Median
17.125 mg/dl
Range
1.2000–100.000 mg/dl
Mean rank
55.62
v2
28.413
39.375
34.333
P
0.000**
0.000**
0.000**
Fig. 1 RT PCR analysis for breast tissue HRG and beta actin RNA by agarose gel electrophoresis and ethidium bromide staining (404, 309 bp, respectively) lane MW molecular weight 100 bp ladder standard (1,000–100 bp), lane 1–4 benign breast conditions showing only positive beta actin bands. Lane 5–6 biopsy from breast cancer patients showing positive HRG and beta actin bands, lane 7, 8: breast biopsy from healthy normal individual showing only positive band for beta actin
expression in breast cancer (Table 1). The sensitivity was 71.7 and 86 % for HRG tissue RNA and HRG serum protein, respectively, while the specificity was 93.3 and 80 %, for HRG RNA and HRG protein, respectively.
Discussion In the current study, we evaluated two novel markers for breast cancer: HRG tissue RNA and HRG serum protein. HRG was initially identified through former studies of breast cancer biomarkers identification in which serum and tissue samples were analyzed using 2-dimensional gel electrophoresis (2-DE) coupled with MS for tissue and
123
Fig. 2 ROC curve analysis for HRG serum protein in malignant versus benign and normal control groups. Area under the curve is 0.837, standard error is 0.056 and confidence limit is (0.727–0.947). Arrow denotes cut off point at 3.1328 mg/dl, at which HRG sensitivity is 86.7 % and specificity, is 80 %
serum analyses. Following manual review of all peptide match spectra and exclusion of redundant matches, HRG was identified that it had sufficient quality to justify the assignment [8, 23]. Schaub et al. [8] declared that MALDITOF proteomic profiling resulted in identification of low mass peptide fragments derived from kininogen, fibrinogen, plasminogen, and intertrypsin inhibitor heavy chain four protein. Adding trypsin digestion indicated increased levels of HRG, prothrombin, and serum amyloid A in patients with late-stage breast cancer. We performed bioinformatics analysis of HRG transcript expression using (GOBO). A prominent finding in GOBO analysis is that the highest levels of HRG expression were found in basal subtypes of breast cancer with
Med Oncol (2014) 31:897
Page 5 of 7 897
Fig. 3 STARD Flowchart summarizing experimental design and the results of this study
very poor OS in the normal-like subtypes, LN- ER? and grade 2 tumor. Patients with intermediate HRG mRNA expression levels have poor survival rate than the patients with significantly high HRG mRNA. The OS rate for the patients with low HRG mRNA levels is high compared with other two categories. In the current study, breast tissue HRG RNA by RT-PCR was significantly expressed in malignant breast tissues compared with fibroadenoma and normal breast tissue. Although the precise molecular mechanisms responsible for HRGmediated tumor activity are still unclear, some researchers assume that HRG inhibits the antiangiogenic effect of the angiogenesis inhibitor TSP-1 and 2 via HRG N terminal and C terminal interfering with binding of TSP to its receptor CD36. Thus, HRG may positively regulate angiogenesis in areas where TSP is abundant. HRG controls vascular smooth muscle cell growth response in pathophysiological states by immobilizing plasminogen/plasmin to the surface of cells, thus enhancing cell migration and invasion [24]. Klenotic et al. [24] demonstrated that the expression of HRG in glioma cells resulted in an increase in brain tumor models tumor size and angiogenesis [25]. No reports in the literatures concerning HRG expression in other human malignancies. In inflammatory conditions, HRG levels in serum changed as a negative acute phase reactant [26]. We failed to find HRG RNA in sera of all investigated groups.
To provide further insight into the role of HRG protein as a novel marker for breast cancer, we measured HRG serum protein by ELISA method. The mean rank of serum HRG was increased more than 1.8-fold in the malignant group compared with the benign group and 2.84-fold compared with the healthy normal group. The positivity rate of serum HRG was significantly correlated with the family history and the grade of the tumor, P \ 0.05. As regards tumor grade, changes in N- and O-linked glycosylation patterns of proteins are commonly observed on the surface of malignant cells and connected to tumor grade and metastasis [27, 28]. The positivity rate of HRG serum protein was highest in moderately differentiated tumors (grade 2) compared with well-differentiated and poorly differentiated tumors in (grades 1, 3); this is consistent with GOBO analysis of HRG transcript and significant expression of HRG mRNA in grade 2. HRG tissue RNA and serum protein were detected in 84.6 and 80.8 %, respectively, in patients with early-stage breast cancers. In the current study, the investigated markers did not show correlation with body mass index, while a significant correlation was observed between HRG protein, menopausal history and OCT intake in agreement with DeMaat et al. [28] who showed that the plasma levels of fibrinogen, PAl-activity, and HRG are influenced by various life style aspects [29]. DeMaat et al.
123
897 Page 6 of 7
Med Oncol (2014) 31:897
Table 2 Sensitivity, specificity, PPV, NPV & accuracy of HRG tissue RNA and HRG serum protein in early (stage I) and in grade 1 breast cancer Parameter
Sensitivity
Specificity
PPV
NPV
Accuracy
Early breast cancer (stage I) (%) HRG tissue RNA
84.6
93.33
91.67
87.5
89.29
HRG serum Protein
80.8
80
77.78
82.7
80.36
Grade 1 breast cancer (%) HRG tissue RNA
83.3
93.33
83.3
93.3
90.47
HRG serum Protein
83.3
80
62.5
92.3
80.95
[28] declared that tumor necrosis alpha and interleukin 1 beta down regulate HRG and mRNA level in primary cultures of hepatocytes. Moreover, scientists reported that oral contraceptive use alters the response pattern of glucocorticoid sensitivity of pro-inflammatory cytokine production to psychosocial stress; thus, they may affect level of HRG transcript [30]. In the malignant group, there was no significant correlation between HRG protein detected by ELISA and HRG RNA measured by RT-PCR (supplementary table 3) since mRNA expression patterns are necessary but are by themselves could be partially correlated with that of the active protein product. This evidence includes posttranscriptional mechanisms including co-factor binding and phosphorylation, the turnover rate of specific proteins or mRNAs, and the intracellular location of the protein products of expressed genes. So mRNA levels cannot be used as substitute for corresponding protein levels without verification which open the way for integrated genomic and proteomic analyses of gene expression [31]. Different immunohistochemical markers have been routinely used by different pathologists to identify basal subtype, but there is no universally agreed-upon set of markers to define this subtype of breast cancer. The expression of HRG in these triple negative subtype may eventually be able to predict which patients will likely respond to a specific molecular therapy and monitor their responses to personalized therapy, since HRG has been referred to as a regulator of angiogenesis, cell adhesion, cell proliferation, and remodeling of the ECM, suggesting that the level of endogenous HRG will influence tumor progression (24).
Conclusions In the current study, the fact that we have adopted strategy to overcome the limitations of the traditional serum
123
proteome comparisons is to use the knowledge about genetic alterations that cause breast cancer to guide the discovery of novel cancer serum biomarkers. Our data revealed that both HRG tissue RNA and serum proteins are novel markers for breast cancer, confirming GOBO analysis. HRG serum protein has a high sensitivity: for detection of low grades and early-stage breast cancers (83.3 and 80.8 %, respectively; Table 2). Further, large multicenter and prospective studies free from bias should be launched for investigating HRG expression as a predictive and prognostic biomarker of clinical outcome which could change practice of progression diagnostics comprehending breast cancer. Acknowledgments This work was supported by the Egyptian Academy of Research and Technology, the Science and Technology Center, Project 21/2. Conflict of interest peting interests.
The authors declare that they have no com-
References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Cancer incidence and mortality worldwide: IARC cancer base no. 10. Lyon: International Agency for Research on Cancer; 2012. 2. Misek DE, Kim EH. Protein biomarkers for the early detection of breast cancer. Int J Proteomics. 2011;2011:343582. 3. Cecchini RS, Costantino JP, Cauley JA, Cronin WM. Body mass index and the risk for developing invasive breast cancer among high-risk women in NSABP P-1 and STAR breast cancer prevention trials. Cancer Prev Res. 2012;5(4):583–92. 4. Cancer Genome Atlas Network. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490(7418):61–70. doi:10.1038/nature11412. 5. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human breast tumors. Nature. 2000;406(6797):747–52. 6. Slamon D, Leyland-Jones B, Shak S, et al. Addition of Herceptin (humanized anti-HER2 antibody) to first line chemotherapy for HER2 overexpressing metastatic breast cancer (HER2?/MBC) markedly increases anticancer activity: a randomized, multinational controlled phase III trial. Presented at: 34th annual meeting of the American Society of Clinical Oncology; May 16–19, 1998. Los Angeles, CA. Abstract 377. 7. Herceptin [prescribing information]. South San Francisco, CA: Genentech, Inc; 2010. 8. Schaub NP, Jones KJ, Nyalwidhe JO, Cazares LH, Karbassi ID, Semmes OJ. Serum proteomic biomarker discovery reflective of stage and obesity in breast cancer patients. J Am Coll Surg. 2009;208(5):970–8. 9. Thulin A. Role of histidine-rich glycoprotein in angiogenesis and tumor growth. Acta Universtatis Upsaliensis. Digital Comprehensive summaries of Uppsala Dissertation from faculty of medicine 505.63 pp. Uppsala; 2009. ISBN-978-91-554-7674-8. 10. Jain S, Zuka M, Liu J, et al. Platelet glycoprotein Ib alpha supports experimental lung metastasis. Proc Natl Acad Sci USA. 2010;104(21):9024–8. 11. Carey LA. Molecular intrinsic subtypes of breast cancer. In Hayes DF, Dizon DS, editors. Up to date. Waltham, MA: Up To Date; 2012.
Med Oncol (2014) 31:897 12. Pinder SE, Elston CW, Ellis IO. In Invasive carcinoma: usual histological types. 3rd ed. In: Elston CW, Ellis IO, editors. The breast. Edinburgh: Churchill Livingstone; 1998. p. 283–337. 13. Sobin L, Gospodarowicz M, Wittekind C, editors. TNM classification of malignant tumors. 7th ed. Hoboken, NJ: Wiley; 2009. 14. World Health Organization. International histological classification of tumours. 2nd ed. Geneva: World Health Organization, 1969–1981. Berlin: Springer; 1988. Present. 15. Hummon AB, Lim SR, Difilippantonio MJ, Ried T. Isolation and solubilization of proteins after TRIzolÒ extraction of RNA and DNA from patient material following prolonged storage. Biotech J. 2007;42:467–72. 16. Koide T, Foster D, Yoshitake S, Davie EW. Amino acid sequence of human histidine rich glycoprotein derived from the nucleotide sequence of its cDNA. Biochem J. 1986;25:2220–5. 17. Sabbatini ARM, Battolla MB, Martini PE, Ranieri-Raggi DM, Moir AJG, Raggi A. Evidence that muscle cells do not express the histidine-rich glycoprotein associated with AMP deaminase but can internalise the plasma protein. Eur J Histochem. 2011;55(1):e6. doi:10.4081/ejh.2011.e6. 18. Smith SD, Wheeler MA, Plescia J, Colberg JW, Weiss RM, Altieri DC. Urine detection of survivin and diagnosis of breast cancer. JAMA J. 2001;285(3):324–8. 19. Zweig MH, Campbell G. Receiver operating characteristic (ROC) Plots, a fundamental evaluation tool in clinical medicine. Clin Chem J. 1993;39(4):561–77. 20. Ringner M, Fredlund E, Hakkinen J, Borg A, Staaf J. GOBO: gene expression-based outcome for breast cancer online. PLoS One J. 2011;6:e17911. 21. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nat J. 2000;406:747–52. 22. Eissa S, Swellam M, Ismail M, Hamdy M, Mokhtar N. Emerging role of p53, bcl-2 and telomerase activity in Egyptian breast cancer patients. IUBMB Life J. 2004;56(8):483–90. 23. Thulin A, Ringvall M, Dimberg A, et al. Activated platelets provide a functional microenvironment for the antiangiogenic
Page 7 of 7 897
24.
25.
26.
27.
28.
29.
30.
31.
fragment of histidine-rich glycoprotein. Mol Cancer Res J. 2009;7(11):1792–802. Klenotic PA, Huang P, Palomo J. Histidine-rich glycoprotein modulates the anti-angiogenic effects of vasculostatin. Am J Pathol. 2010;176(4):2039–50. Saigo K, Yoshida A, Sugano W, Ryo R, Yamaguchi N. Histidinerich glycoprotein in blood during inflammation, surgical operation or hemodialysis. Rinsho Ketsueki. 1990;31(12):1914–9. Menga R, Gormleya M, Vadiraja B, Rosenbergc BA, Quonga AA. Low abundance protein enrichment for discovery of candidate plasma protein biomarkers for early detection of breast cancer. J Proteomics. 2011;1(75):366–74. Reeves GK, Beral V, Green J, Gathani T, Bull D. Hormonal therapy for menopause and breast cancer risk by histological type: a cohort study and meta-analysis. Lancet Oncol J. 2006;7:910–8. DeMaat MPM, Kameling SWA, Kluft C. The sensitivity and specificity of some clotting rate and immunological fibrinogen assays for high, low and low molecular weight forms of fibrinogen. In: DeMaat MPM, editor. Regulation and modulation of the plasma fibrinogen level. Rotterdam: University of Rotterdam. Thesis; 1995. pp. 47–59. Rohleder N, Wolf JM, Piel M, Kirschbaum C. Impact of oral contraceptive use on glucocorticoid sensitivity of pro-inflammatory cytokine production after psychosocial stress. Psychoneuroendocrinology. 2003;28(3):261–73. Schwanhausser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, Chen W, Selbach M. Global quantification of mammalian gene expression control. Nature. 2011;473(7347):337–42. Shebl F, Pinto LA, Garcia-Pineres A, Lempicki R, Williams M, Harro C, Hildesheim A. Comparison of mRNA and protein measures of cytokines following vaccination with human papillomavirus-16 L1 virus-like particles. Cancer Epidemiol Biomarkers Prev. 2010;19(4):978–81.
123
ORIGINAL PAPER
Evaluation of histidine-rich glycoprotein tissue RNA and serum protein as novel markers for breast cancer Marwa Matboli • Sanaa Eissa • Hebatallah Said
Received: 4 October 2013 / Accepted: 17 February 2014 Ó Springer Science+Business Media New York 2014
Abstract Advances in the field of breast cancer (BC) biomarkers discovery facilitate diagnosis and treatment of BC in its pre-invasive state. While the genetic tissue markers are making significant advances in understanding the molecular basis of BC, serum has long been considered a rich source for biomarkers. So, integrated genomic and proteomic strategies play a huge role in the analytical validation of BC biomarkers and represent a true milestone in the areas of diagnostics and personalized medicine. This study included 60 cases (BC), 30 patients with fibroadenoma and 30 healthy women. Histidine-rich glycoprotein RNA (HRG) tissue expression was analyzed through gene expression-based outcome for breast cancer online algorithm (GOBO) bioinformatic analysis. To confirm our informatics analysis, HRG RNA was detected in breast tissue samples by RT-PCR, and HRG serum protein was estimated by ELISA. GOBO analysis revealed increased HRG RNA expression in all subtypes of BC with relative higher expression in basal subtype and grade 2. We confirmed these findings by HRG tissue RNA with 71.7 % sensitivity and 93.3 % specificity. HRG serum protein was 86.7 % sensitivity and 80 % specificity. HRG tissue RNA and serum protein could be considered as promising novel markers for prediction of BC prognosis.
Electronic supplementary material The online version of this article (doi:10.1007/s12032-014-0897-4) contains supplementary material, which is available to authorized users. M. Matboli S. Eissa (&) H. Said Oncology Diagnostic Unit, Medical Biochemistry and Molecular Biology Department, Faculty of Medicine, Ain Shams University, P.O. Box 11381, Abbassia, Cairo, Egypt e-mail: [email protected]; [email protected]; [email protected]
Keywords Breast carcinoma Bioinformatics Histidinerich glycoprotein ELISA Serum markers Abbreviations HRG Histidine-rich glycoprotein BC Breast cancer BMI Body mass index LN Lymph node ER Estrogen receptor PR Progesterone receptor Her-2 neu Human epidermal growth factor receptor 2 HT Hormonal therapy OCT Oral contraceptive therapy
Introduction Breast cancer is by far the most common cancer among women of both developed and developing countries accounting for 22.9 % of all female cancers. It is also the main cause of cancer death in females responsible for 13.7 % of their cancer-related mortality. In Egypt, breast cancer is estimated to be the most common cancer among females accounting for 37.7 % of their total cancer cases. It is also the major cause of cancer-related mortality accounting for 29.1 % of their cancer-related deaths [1]. The most efficacious screening tool utilized in the clinic is mammography, leading to great improvement in mortality reduction though lesions \0.5 cm in size remains undetectable by such technology [2]. There is also resistance for seeking such services yearly; many false positives are encountered, and cost-to-benefit ratio of these imaging methodologies is high [3]. Breast cancer subtyping has turned out to be the golden era toward new therapeutic strategies. For example, breast
123
897 Page 2 of 7
cancer basal subtype is characterized by high histologic grade, the absence of HER-2 (ErB2) and receptors for estrogen and progesterone with no approved targeted therapies yet exist, since they do not express any known target proteins [4–7]. Histidine-rich glycoprotein was initially identified through former proteomic studies for breast cancer biomarkers identification [8]. HRG is a2-glycoprotein synthesized by liver and present in plasma and platelets. HRG is potentially involved in numerous biologic processes. It can inhibit rosette formation and interacts with heparin, thrombospondin and plasminogen. There have been conflicting reports on both enhancing and inhibitory effects of HRG on angiogenesis [9, 10]. In this study, we explore a two-stage strategy for evaluation of a serum biomarker corresponding to specific cancer-associated gene over-expression and its application in breast cancer.
Med Oncol (2014) 31:897
who underwent plastic breast surgery (mean age 50.47 ± 11.64 years and range from 36 to 77 years) with matching age and sex to the patient’ groups. Study population demographic and clinical characteristics are shown in (Supplementary table 1). Sample processing Sera were collected before surgery following a standardized protocol and stored within 15 min at -80 °C till sample preparation. A smaller volume aliquots are initially prepared to minimize effect of freeze–thaw cycles and provide multiple replicates that were processed in an identical manner. After primary surgery, a representative part of the tumor tissue was immediately frozen stored at -80 °C until RNA extraction. The samples were coded, and clinical information was unavailable to the technicians performing the mRNA quantification. Bioinformatics analysis of HRG
Materials and methods Study population (Supplementary table 1) This pilot study was approved by Ain Shams Faculty of Medicine ethical committee. A prospective analysis was performed on 120 Egyptian subjects, and 90 patients were selected from The General Surgery department, Ain Shams University Hospitals, Egypt, between November 2010 and June 2012, after obtaining informed consent. Patients presented with breast mass and/or bleeding per nipple were included in the study, and patients with past history of breast cancer or any other malignancy or inflammatory mastitis were excluded from the study. All patients provided paired tissue and blood samples. Breast cancer specimens were obtained after modified radical mastectomy for operable primary breast cancer. The pathology specimens were examined independently by histopathologists to grade and subclassify the tumors according to established criteria [11, 12]. Subsequently, the 90 patients included in the study were classified into malignant and benign groups. The malignant group included 60 patients (mean age 52.47 ± 13.2 years and range from 20 to 81 years). Of those patients, 43 were diagnosed by histopathology with invasive duct carcinoma (IDC), 1 patient with invasive lobular carcinoma (ILC), 9 with mixed invasive duct and lobular carcinoma, and 7 patients with rare histological types. Tumor staging and grading were determined according to TNM and World Health Organization classification [13, 14]. The benign group included 30 patients with benign breast diseases: fibroadenoma (49.6 ± 14.11 years and range from 20 to 67 years). The third study group was represented by 30 healthy volunteers
123
Histidine-rich glycoprotein was selected from a large set of signature peptides identified by Schaub et al. [8]. This includes the fragments with the highest significant relation to breast cancer, followed by bioinformatic analysis of HRG transcript expression using data from the gene expression-based outcome for breast cancer online algorithm (GOBO). Sample processing Sera and breast tissues were collected following a standardized protocol (supplementary files). Total RNA was purified with TriFastTM (Pe QLab Biotechnologie GmbH Corporation) according to the manufacturer’s instructions [15]. RT-PCR for HRG RNA RT-PCRs were done in Hybaid thermal cycler [Thermo Electron (formerly Hybaid) Waltham, MA, USA] using Qiagen one-step RT-PCR Kit (Cat. No. 210210) using HRG primers [16–18] (accession NM_000412.2) (more details on supplementary files), yielding an amplicon of 404 bp separated on 2 % agarose gel stained with ethidium bromide. The quality of RNA was checked by amplification of c-DNA for b-actin to ensure successful completion of c-DNA using primer pair [19] (Accession NM_001017992.2), yielding an amplicon of 309 bp (Fig. 2, supplementary figure 3). HRG protein quantitative estimation Sera from all subjects were subjected to immunoassay to measure serum level of HRG protein by using Enzyme-
Med Oncol (2014) 31:897
linked immunosorbent assay kit for HRG supplied by USCN, Life Science Inc., (China). Statistical analysis Univariate analyses were performed using a chi-square test of association of categoric variables. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated according to standard statistical methods. All analyses were performed using Statistical Package for the Social Sciences software (SPSS Inc., Chicago, IL, USA). The threshold value for optimal sensitivity and specificity of HRG was determined by receiver operating characteristics (ROC) curve [20].
Page 3 of 7 897
breast cancer cases (Table 1; Fig. 1 and supplementary figure 3). There was a statistically significant difference in the expression of HRG among the three groups (P = 0.000) (Table 1). No significant correlation was found between HRG mRNA positivity rate and any of the studied clinicopathological factors except molecular subtype and grade (P [ 0.05) (Supplementary table 2). Results demonstrated no association with ER? tumors (55.8 %) and a higher association with basal subtypes (100 %) and grade 2 tumors (68 %). These data are consistent with the correlations recorded in the GOBO data (Supplementary Figure 1a,b). On the other hand, we failed to find HRG mRNA in all investigated serum samples using RT-PCR. Serum HRG protein analysis
Results HRG transcript expression GOBO bioinformatic analysis Bioinformatics analysis of HRG transcript expression was assessed using data from GOBO [21]. GOBO is a web-based analysis tool that utilizes Affymetrix gene expression data collected from 1,881 breast cancer patients with associated stage, grade, nodal status and intrinsic molecular classification based on the paradigm first reported by the Perou Laboratory [22]). Expression of HRG showed statistically significant difference between different subtypes (P \ 0. 001) with higher expression in basal subtypes than other subtypes. Moreover, significant HRG RNA expression in relation to grade (P = 0.00011) with higher expression in grade 2 tumors were compared with grade 1 and 3 (Supplementary figure 1a-c). The lowest HRG transcript expression was observed in HER2-enriched. We performed through GOBO tool a series of Kaplan–Meier analyses to determine whether HRG expression is associated with overall survival (OS). Positive tissue HRG RNA expression is associated with survival when considering all breast tumor subtypes together, and it is highly associated with poor OS in normal like (P = 0.0023), Lymph node– ER? tumors (P = 0.004888) and grade 2 (P = 0.0079) (Supplementary figure 2a-e). Breast tissue HRG mRNA expression Results from the GOBO transcript expression analysis motivated our research group to investigate HRG mRNA at the gene level in addition to the protein level in sera of breast cancer patients. HRG was validated by performing RT-PCR (reverse transcription polymerase chain reaction) for HRG mRNA expression in the breast biopsies. HRG mRNA was not detected in 30 normal breast tissues, but was detected in 13.3 % (4/30) of fibroadenoma cases and 71.7 % (43/60) of
Histidine-rich glycoprotein protein level was significantly higher in the malignant group (Mean rank = 55. 62) as compared to benign (Mean rank = 30.93) and normal control groups (Mean rank = 19.60), P \ 0.001; ROC curve was constructed to determine the best cutoff value of HRG protein that discriminates malignant from non-malignant groups (Fig. 2). The cutoff value was set at 3.1 mg/dl, and the positivity rate of serum HRG protein (No of cases [3.1 mg/dl) was estimated accordingly among all the study groups (Table 1). In the healthy control group, HRG serum protein positivity rate was (13.3 %), in benign group (26.3 %), while the malignant group expressed the highest positivity rate (86.7 %) with a highly statistically significant difference between the three groups (P = 0.000). No significant correlation was found between HRG protein positivity rate and any of the studied clinicopathological factors except grade, menopausal history, and oral contraceptive treatment (OCT) P [ 0.05 (Supplementary table 2, Fig. 3). HRG serum protein in different breast cancer molecular subtypes HRG serum protein was not associated with ER? tumors, but was highly associated with basal subtypes (all of the basal subtypes show over-expression of serum HRG protein 30/30 (Supplementary Figure 1a,b, Supplementary table 2). These data are consistent with the correlation observed in the GOBO data. Interestingly, grade 2 tumors were associated with a higher serum HRG protein positivity (Supplementary Figure 1c). Sensitivity, specificity, positive predictive value, negative predictive value and accuracy of investigated parameters Sensitivity, specificity, PPV, NPV and accuracy of HRG tissue RNA, and serum HRG protein were tested for their
123
897 Page 4 of 7 Table 1 HRG tissue RNA and serum protein level and positivity rate in the malignant group compared with benign and normal control groups
Med Oncol (2014) 31:897
Group
Serum HRG level
No. of HRG protein [3.1328 mg/dl (%)
No. of HRG RNA positive in breast tissue (%)
Median
1.1333 mg/dl
4/30 (13.3 %)
0/30 (0 %)
Range
0.7107-52.9000 mg/dl
Mean rank
19.6 8/30 (26.7 %)
4/30 (13.3 %)
52/60 (86.7 %)
43/60 (71.7 %)
Normal control (n = 30)
Benign (n = 30) Median
1.5600 mg/dl
Range Mean rank
0.7107–52.9000 mg/dl 30.93
Malignant (n = 60) Median
17.125 mg/dl
Range
1.2000–100.000 mg/dl
Mean rank
55.62
v2
28.413
39.375
34.333
P
0.000**
0.000**
0.000**
Fig. 1 RT PCR analysis for breast tissue HRG and beta actin RNA by agarose gel electrophoresis and ethidium bromide staining (404, 309 bp, respectively) lane MW molecular weight 100 bp ladder standard (1,000–100 bp), lane 1–4 benign breast conditions showing only positive beta actin bands. Lane 5–6 biopsy from breast cancer patients showing positive HRG and beta actin bands, lane 7, 8: breast biopsy from healthy normal individual showing only positive band for beta actin
expression in breast cancer (Table 1). The sensitivity was 71.7 and 86 % for HRG tissue RNA and HRG serum protein, respectively, while the specificity was 93.3 and 80 %, for HRG RNA and HRG protein, respectively.
Discussion In the current study, we evaluated two novel markers for breast cancer: HRG tissue RNA and HRG serum protein. HRG was initially identified through former studies of breast cancer biomarkers identification in which serum and tissue samples were analyzed using 2-dimensional gel electrophoresis (2-DE) coupled with MS for tissue and
123
Fig. 2 ROC curve analysis for HRG serum protein in malignant versus benign and normal control groups. Area under the curve is 0.837, standard error is 0.056 and confidence limit is (0.727–0.947). Arrow denotes cut off point at 3.1328 mg/dl, at which HRG sensitivity is 86.7 % and specificity, is 80 %
serum analyses. Following manual review of all peptide match spectra and exclusion of redundant matches, HRG was identified that it had sufficient quality to justify the assignment [8, 23]. Schaub et al. [8] declared that MALDITOF proteomic profiling resulted in identification of low mass peptide fragments derived from kininogen, fibrinogen, plasminogen, and intertrypsin inhibitor heavy chain four protein. Adding trypsin digestion indicated increased levels of HRG, prothrombin, and serum amyloid A in patients with late-stage breast cancer. We performed bioinformatics analysis of HRG transcript expression using (GOBO). A prominent finding in GOBO analysis is that the highest levels of HRG expression were found in basal subtypes of breast cancer with
Med Oncol (2014) 31:897
Page 5 of 7 897
Fig. 3 STARD Flowchart summarizing experimental design and the results of this study
very poor OS in the normal-like subtypes, LN- ER? and grade 2 tumor. Patients with intermediate HRG mRNA expression levels have poor survival rate than the patients with significantly high HRG mRNA. The OS rate for the patients with low HRG mRNA levels is high compared with other two categories. In the current study, breast tissue HRG RNA by RT-PCR was significantly expressed in malignant breast tissues compared with fibroadenoma and normal breast tissue. Although the precise molecular mechanisms responsible for HRGmediated tumor activity are still unclear, some researchers assume that HRG inhibits the antiangiogenic effect of the angiogenesis inhibitor TSP-1 and 2 via HRG N terminal and C terminal interfering with binding of TSP to its receptor CD36. Thus, HRG may positively regulate angiogenesis in areas where TSP is abundant. HRG controls vascular smooth muscle cell growth response in pathophysiological states by immobilizing plasminogen/plasmin to the surface of cells, thus enhancing cell migration and invasion [24]. Klenotic et al. [24] demonstrated that the expression of HRG in glioma cells resulted in an increase in brain tumor models tumor size and angiogenesis [25]. No reports in the literatures concerning HRG expression in other human malignancies. In inflammatory conditions, HRG levels in serum changed as a negative acute phase reactant [26]. We failed to find HRG RNA in sera of all investigated groups.
To provide further insight into the role of HRG protein as a novel marker for breast cancer, we measured HRG serum protein by ELISA method. The mean rank of serum HRG was increased more than 1.8-fold in the malignant group compared with the benign group and 2.84-fold compared with the healthy normal group. The positivity rate of serum HRG was significantly correlated with the family history and the grade of the tumor, P \ 0.05. As regards tumor grade, changes in N- and O-linked glycosylation patterns of proteins are commonly observed on the surface of malignant cells and connected to tumor grade and metastasis [27, 28]. The positivity rate of HRG serum protein was highest in moderately differentiated tumors (grade 2) compared with well-differentiated and poorly differentiated tumors in (grades 1, 3); this is consistent with GOBO analysis of HRG transcript and significant expression of HRG mRNA in grade 2. HRG tissue RNA and serum protein were detected in 84.6 and 80.8 %, respectively, in patients with early-stage breast cancers. In the current study, the investigated markers did not show correlation with body mass index, while a significant correlation was observed between HRG protein, menopausal history and OCT intake in agreement with DeMaat et al. [28] who showed that the plasma levels of fibrinogen, PAl-activity, and HRG are influenced by various life style aspects [29]. DeMaat et al.
123
897 Page 6 of 7
Med Oncol (2014) 31:897
Table 2 Sensitivity, specificity, PPV, NPV & accuracy of HRG tissue RNA and HRG serum protein in early (stage I) and in grade 1 breast cancer Parameter
Sensitivity
Specificity
PPV
NPV
Accuracy
Early breast cancer (stage I) (%) HRG tissue RNA
84.6
93.33
91.67
87.5
89.29
HRG serum Protein
80.8
80
77.78
82.7
80.36
Grade 1 breast cancer (%) HRG tissue RNA
83.3
93.33
83.3
93.3
90.47
HRG serum Protein
83.3
80
62.5
92.3
80.95
[28] declared that tumor necrosis alpha and interleukin 1 beta down regulate HRG and mRNA level in primary cultures of hepatocytes. Moreover, scientists reported that oral contraceptive use alters the response pattern of glucocorticoid sensitivity of pro-inflammatory cytokine production to psychosocial stress; thus, they may affect level of HRG transcript [30]. In the malignant group, there was no significant correlation between HRG protein detected by ELISA and HRG RNA measured by RT-PCR (supplementary table 3) since mRNA expression patterns are necessary but are by themselves could be partially correlated with that of the active protein product. This evidence includes posttranscriptional mechanisms including co-factor binding and phosphorylation, the turnover rate of specific proteins or mRNAs, and the intracellular location of the protein products of expressed genes. So mRNA levels cannot be used as substitute for corresponding protein levels without verification which open the way for integrated genomic and proteomic analyses of gene expression [31]. Different immunohistochemical markers have been routinely used by different pathologists to identify basal subtype, but there is no universally agreed-upon set of markers to define this subtype of breast cancer. The expression of HRG in these triple negative subtype may eventually be able to predict which patients will likely respond to a specific molecular therapy and monitor their responses to personalized therapy, since HRG has been referred to as a regulator of angiogenesis, cell adhesion, cell proliferation, and remodeling of the ECM, suggesting that the level of endogenous HRG will influence tumor progression (24).
Conclusions In the current study, the fact that we have adopted strategy to overcome the limitations of the traditional serum
123
proteome comparisons is to use the knowledge about genetic alterations that cause breast cancer to guide the discovery of novel cancer serum biomarkers. Our data revealed that both HRG tissue RNA and serum proteins are novel markers for breast cancer, confirming GOBO analysis. HRG serum protein has a high sensitivity: for detection of low grades and early-stage breast cancers (83.3 and 80.8 %, respectively; Table 2). Further, large multicenter and prospective studies free from bias should be launched for investigating HRG expression as a predictive and prognostic biomarker of clinical outcome which could change practice of progression diagnostics comprehending breast cancer. Acknowledgments This work was supported by the Egyptian Academy of Research and Technology, the Science and Technology Center, Project 21/2. Conflict of interest peting interests.
The authors declare that they have no com-
References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Cancer incidence and mortality worldwide: IARC cancer base no. 10. Lyon: International Agency for Research on Cancer; 2012. 2. Misek DE, Kim EH. Protein biomarkers for the early detection of breast cancer. Int J Proteomics. 2011;2011:343582. 3. Cecchini RS, Costantino JP, Cauley JA, Cronin WM. Body mass index and the risk for developing invasive breast cancer among high-risk women in NSABP P-1 and STAR breast cancer prevention trials. Cancer Prev Res. 2012;5(4):583–92. 4. Cancer Genome Atlas Network. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490(7418):61–70. doi:10.1038/nature11412. 5. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human breast tumors. Nature. 2000;406(6797):747–52. 6. Slamon D, Leyland-Jones B, Shak S, et al. Addition of Herceptin (humanized anti-HER2 antibody) to first line chemotherapy for HER2 overexpressing metastatic breast cancer (HER2?/MBC) markedly increases anticancer activity: a randomized, multinational controlled phase III trial. Presented at: 34th annual meeting of the American Society of Clinical Oncology; May 16–19, 1998. Los Angeles, CA. Abstract 377. 7. Herceptin [prescribing information]. South San Francisco, CA: Genentech, Inc; 2010. 8. Schaub NP, Jones KJ, Nyalwidhe JO, Cazares LH, Karbassi ID, Semmes OJ. Serum proteomic biomarker discovery reflective of stage and obesity in breast cancer patients. J Am Coll Surg. 2009;208(5):970–8. 9. Thulin A. Role of histidine-rich glycoprotein in angiogenesis and tumor growth. Acta Universtatis Upsaliensis. Digital Comprehensive summaries of Uppsala Dissertation from faculty of medicine 505.63 pp. Uppsala; 2009. ISBN-978-91-554-7674-8. 10. Jain S, Zuka M, Liu J, et al. Platelet glycoprotein Ib alpha supports experimental lung metastasis. Proc Natl Acad Sci USA. 2010;104(21):9024–8. 11. Carey LA. Molecular intrinsic subtypes of breast cancer. In Hayes DF, Dizon DS, editors. Up to date. Waltham, MA: Up To Date; 2012.
Med Oncol (2014) 31:897 12. Pinder SE, Elston CW, Ellis IO. In Invasive carcinoma: usual histological types. 3rd ed. In: Elston CW, Ellis IO, editors. The breast. Edinburgh: Churchill Livingstone; 1998. p. 283–337. 13. Sobin L, Gospodarowicz M, Wittekind C, editors. TNM classification of malignant tumors. 7th ed. Hoboken, NJ: Wiley; 2009. 14. World Health Organization. International histological classification of tumours. 2nd ed. Geneva: World Health Organization, 1969–1981. Berlin: Springer; 1988. Present. 15. Hummon AB, Lim SR, Difilippantonio MJ, Ried T. Isolation and solubilization of proteins after TRIzolÒ extraction of RNA and DNA from patient material following prolonged storage. Biotech J. 2007;42:467–72. 16. Koide T, Foster D, Yoshitake S, Davie EW. Amino acid sequence of human histidine rich glycoprotein derived from the nucleotide sequence of its cDNA. Biochem J. 1986;25:2220–5. 17. Sabbatini ARM, Battolla MB, Martini PE, Ranieri-Raggi DM, Moir AJG, Raggi A. Evidence that muscle cells do not express the histidine-rich glycoprotein associated with AMP deaminase but can internalise the plasma protein. Eur J Histochem. 2011;55(1):e6. doi:10.4081/ejh.2011.e6. 18. Smith SD, Wheeler MA, Plescia J, Colberg JW, Weiss RM, Altieri DC. Urine detection of survivin and diagnosis of breast cancer. JAMA J. 2001;285(3):324–8. 19. Zweig MH, Campbell G. Receiver operating characteristic (ROC) Plots, a fundamental evaluation tool in clinical medicine. Clin Chem J. 1993;39(4):561–77. 20. Ringner M, Fredlund E, Hakkinen J, Borg A, Staaf J. GOBO: gene expression-based outcome for breast cancer online. PLoS One J. 2011;6:e17911. 21. Perou CM, Sørlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nat J. 2000;406:747–52. 22. Eissa S, Swellam M, Ismail M, Hamdy M, Mokhtar N. Emerging role of p53, bcl-2 and telomerase activity in Egyptian breast cancer patients. IUBMB Life J. 2004;56(8):483–90. 23. Thulin A, Ringvall M, Dimberg A, et al. Activated platelets provide a functional microenvironment for the antiangiogenic
Page 7 of 7 897
24.
25.
26.
27.
28.
29.
30.
31.
fragment of histidine-rich glycoprotein. Mol Cancer Res J. 2009;7(11):1792–802. Klenotic PA, Huang P, Palomo J. Histidine-rich glycoprotein modulates the anti-angiogenic effects of vasculostatin. Am J Pathol. 2010;176(4):2039–50. Saigo K, Yoshida A, Sugano W, Ryo R, Yamaguchi N. Histidinerich glycoprotein in blood during inflammation, surgical operation or hemodialysis. Rinsho Ketsueki. 1990;31(12):1914–9. Menga R, Gormleya M, Vadiraja B, Rosenbergc BA, Quonga AA. Low abundance protein enrichment for discovery of candidate plasma protein biomarkers for early detection of breast cancer. J Proteomics. 2011;1(75):366–74. Reeves GK, Beral V, Green J, Gathani T, Bull D. Hormonal therapy for menopause and breast cancer risk by histological type: a cohort study and meta-analysis. Lancet Oncol J. 2006;7:910–8. DeMaat MPM, Kameling SWA, Kluft C. The sensitivity and specificity of some clotting rate and immunological fibrinogen assays for high, low and low molecular weight forms of fibrinogen. In: DeMaat MPM, editor. Regulation and modulation of the plasma fibrinogen level. Rotterdam: University of Rotterdam. Thesis; 1995. pp. 47–59. Rohleder N, Wolf JM, Piel M, Kirschbaum C. Impact of oral contraceptive use on glucocorticoid sensitivity of pro-inflammatory cytokine production after psychosocial stress. Psychoneuroendocrinology. 2003;28(3):261–73. Schwanhausser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, Chen W, Selbach M. Global quantification of mammalian gene expression control. Nature. 2011;473(7347):337–42. Shebl F, Pinto LA, Garcia-Pineres A, Lempicki R, Williams M, Harro C, Hildesheim A. Comparison of mRNA and protein measures of cytokines following vaccination with human papillomavirus-16 L1 virus-like particles. Cancer Epidemiol Biomarkers Prev. 2010;19(4):978–81.
123
