UBC Faculty Research and Publications

Redefining prognostic factors for breast cancer: YB-1 is a stronger predictor of relapse and disease-specific… Habibi, Golareh; Leung, Samuel; Law, Jennifer H; Gelmon, Karen; Masoudi, Hamid; Turbin, Dmitry; Pollak, Michael; Nielsen, Torsten O; Huntsman, David; Dunn, Sandra E Oct 16, 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13058_2008_Article_2120.pdf [ 953.16kB ]
JSON: 52383-1.0220541.json
JSON-LD: 52383-1.0220541-ld.json
RDF/XML (Pretty): 52383-1.0220541-rdf.xml
RDF/JSON: 52383-1.0220541-rdf.json
Turtle: 52383-1.0220541-turtle.txt
N-Triples: 52383-1.0220541-rdf-ntriples.txt
Original Record: 52383-1.0220541-source.json
Full Text

Full Text

Available online http://breast-cancer-research.com/content/10/5/R86Open AccessVol 10 No 5Research articleRedefining prognostic factors for breast cancer: YB-1 is a stronger predictor of relapse and disease-specific survival than estrogen receptor or HER-2 across all tumor subtypesGolareh Habibi1, Samuel Leung2, Jennifer H Law1, Karen Gelmon3, Hamid Masoudi4, Dmitry Turbin2, Michael Pollak5, Torsten O Nielsen2, David Huntsman2,4 and Sandra E Dunn11Laboratory for Oncogenomic Research, Departments of Pediatrics and Experimental Medicine, Child and Family Research Institute, W. 28th Avenue, University of British Columbia, Vancouver, BC, V5Z 4H4, Canada2Genetic Pathology Evaluation Center, Jack Bell Research Center, Oak Street, Vancouver, BC, V6H 3Z6, Canada3Division of Medical Oncology, British Columbia Cancer Agency, West 10th Avenue, Vancouver, BC, V5Z 1L3 Canada4Department of Pathology and Laboratory, West 10th Avenue, Vancouver, BC, V5Z 1L3, Canada5Division of Medical Oncology, Jewish General Hospital – Lady Davis Institute, Côte Ste Catherine Road, Montreal, Quebec H3T 1E2, CanadaCorresponding author: Sandra E Dunn, sedunn@interchange.ubc.caReceived: 11 Jun 2008 Revisions requested: 11 Jul 2008 Revisions received: 17 Sep 2008 Accepted: 16 Oct 2008 Published: 16 Oct 2008Breast Cancer Research 2008, 10:R86 (doi:10.1186/bcr2156)This article is online at: http://breast-cancer-research.com/content/10/5/R86© 2008 Habibi et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.AbstractIntroduction Gene expression analysis is used to subtypebreast cancers such that the most aggressive tumors areidentified, but translating this into clinical practice can becumbersome. Our goal is to develop a universal biomarker thatdistinguishes patients at high risk across all breast cancersubtypes. We previously reported that Y-box binding protein-1(YB-1), a transcription/translation factor, was a marker of poorprognosis in a cohort of 490 patients with breast cancer, but thestudy was not large enough to subtype the cancers. Wetherefore investigated whether YB-1 identifies patients at risk foreither reduced relapse free survival or decreased r breastcancer specific survival (BCSS) across all tumor subtypes byevaluating 4,049 cases.Methods Tumor tissue microarrays, representing 4,049 casesof invasive breast cancers with 20 years of follow up, weresubtyped by the expression profiles of estrogen receptor,progesterone receptor, or HER-2. We then addressed whetherYB-1 expression identified patients at higher risk for relapseand/or lower BCSS.Results We found YB-1 to be a highly predictive biomarker ofrelapse (P < 2.5 × 10-20) and poor survival (P < 7.3 × 10-26) inthe entire cohort and across all breast cancer subtypes. Patientswith node-positive or node-negative cancer were more likely todie from the disease if YB-1 was expressed. This was furthersubstantiated using a Cox regression model, which revealedthat it was significantly associated with relapse and poor survivalin a subtype independent manner (relapse patients, hazard ratio= 1.28, P < 8 × 10-3; all patients, hazard ratio = 1.45, P < 6.7× 10-7). Moreover, YB-1 was superior to estrogen receptor andHER-2 as a prognostic marker for relapse and survival. For asubset of patients who were originally considered low risk andwere therefore not given chemotherapy, YB-1 was indicative ofpoor survival (P < 7.1 × 10 -17). Likewise, YB-1 was predictiveof decreased BCSS in tamoxifen-treated patients (P = 0.001);in this setting a Cox regression model once again demonstratedit to be an independent biomarker indicating poor survival(hazard ratio = 1.70, P = 0.022).Conclusions Expression of YB-1 universally identifies patientsat high risk across all breast cancer subtypes and in situationswhere more aggressive treatment may be needed. We thereforepropose that YB-1 may re-define high-risk breast cancer andthereby create opportunities for individualized therapy.IntroductionThe overall goal of predictive oncology is to refine treatmentoptions for patients such that they may receive optimal carewithout experiencing unnecessary side effects. One of thePage 1 of 9(page number not for citation purposes)BCSS: breast cancer specific survival; CI: confidence interval; EGFR: epidermal growth factor receptor; ER: estrogen receptor; HER: human epider-mal growth factor receptor; HR: hazard ratio; RT-PCR: reverse transcription polymerase chain reaction; PR: progesterone receptor; RFS: relapse free survival; TMA: tissue microarray; YB-1: Y-box binding protein-1.Breast Cancer Research    Vol 10 No 5    Habibi et al.greatest challenges is the identification and implementation ofbiomarkers for cancer [1]. Although the antigen Ki67 is report-edly associated with poor survival it is no longer recom-mended as a biomarker in prognostic groups, according to arecent report from the American Society for Clinical Oncology[2]. Furthermore, this touted biomarker was disappointinglynotpredictive of response to adjuvant chemoendocrine ther-apy in a study of about 2,000 patients enrolled in two rand-omized International Breast Cancer Study Group trials [3].Many reports have shown that human epidermal growth factorreceptor (HER)-2 is a marker of poor prognosis in breast can-cer, following its initial report in 1987 by Slamon and cowork-ers [4]. This subsequently led to the development of targetedagents against it. However, where it is most clinically useful isas a predictive marker used to guide treatment decisionsabout whether to use agents that target this receptor such astrastuzumab [3]. In fact, HER-2 is no longer recommended asa prognostic factor in breast cancer [3]. In contrast to HER-2,the estrogen receptor (ER) is associated with good prognosis[5,6]. ER is similar to HER-2 in that many therapies targetedagainst it have been developed. As such, ER is more routinelyused as a biomarker to guide treatment decisions aboutwhether hormone therapy is appropriate [3]. Although HER-2and ER have been very informative for our understanding ofpatient survival, and this has lead to the eventual developmentof targeted agents, they do not apply to all tumor typesGiventhe limited number of robust biomarkers that predict poor over-all survival, the question of which should be used to guidepatient care remains open [7]. Breast cancers have been sub-divided into four subtypes, namely luminal A, luminal B, HER-2, and basal-like (also called triple negative), based on geneexpression signature [8]. The latter two types are typicallymore aggressive than the former two. It should be noted that abiomarker denoting poor survival across subtypes is yet to beidentified.Y-box binding protein-1 (YB-1) is a transcription and transla-tion factor that can promote tumor growth and chemotherapyresistance by inducing growth-promoting genes such as HER-2 and EGFR (epidermal grwoth factor receptor) [9], PCNA(proliferating cell nuclear antigen) [10], MDR-1/ABCA1(multi-drug resistance-1) [11], cyclin A [12], and cyclin B [12].YB-1 (also known as nuclease sensitive protein-1 or NSEP-1)was reportedly expressed in a high-risk group of patients withBRCA1 mutations by cDNA microarray [13]. Perou and Sorliealso identified it by microarray analysis [8], but at the transcriptlevel YB-1 clustered into a unique group of genes, the signifi-cance of which awaits further description. Perhaps this signa-ture represents genes that are commonly over-expressed inbreast cancer but do not to fall into a given subtype. Withregard to tumor progression, targeted expression of YB-1 inthe mammary gland under the WAP promoter leads to thedevelopment of tumors; 100% of mice develop invasive carci-These studies indicate that YB-1 is indeed an oncogene thatis important in the genesis of the disease [14]. In models ofhuman breast cancer, inhibition of YB-1 with a dominant neg-ative mutant (interferes with the DNA binding activity of theprotein [S102A]) slows tumor cell growth [15], and this isassociated with the downregulation of EGFR and HER-2 [9].It thus appears that YB-1 is important in the development ofmammary tumors and that human breast cancer cells not onlyexpress this oncoprotein but also continue to depend upon itfor sustained growth and survival.It is unclear whether YB-1 carries prognostic value for specifictypes of breast cancer or whether it may be a useful biomarkeracross all types. In a pilot study, we reported that YB-1 expres-sion is associated with poor overall survival in a group of 490patients with invasive breast cancer [9]. In that cohort we alsofound that YB-1 was expressed in more than 70% of basal-likebreast cancers [16], but these studies were limited by inade-quate size to determine patient survival within individual sub-types. We also lacked the power to evaluate the importantquestion of whether YB-1 expression is associated with higherrates of relapse. To address these issues, we screened a largetissue microarray (TMA) series consisting of 4,049 invasivebreast cancers with 20 years of clinical follow up. The potentialvalue of YB-1 as a biomarker for aggressive disease was alsoexamined in a subset of women who were originally consid-ered to be at low risk and therefore did not receive chemother-apy. Finally, we examined the possibility that YB-1 may beassociated with more aggressive disease in an adjuvant set-ting in which tamoxifen was given for 5 years.Materials and methodsThe study cohort and construction of the TMAs used in thepresent study has previously been described [5,17]. Sectionsfrom were cut at 4 μm and immunostained with a rabbit poly-clonal anti-YB-1 antibody (1:1,400; a gift from Dr Colleen Nel-son, University of British Columbia, Vancouver, BC, Canada).Of note, this antibody detects marked YB-1 expression in thecytoplasm of formalin-fixed paraffin-embedded tissues, butidentification of YB-1 in the nucleus is much lower than onewould expect. We believe that this is because the carboxyl-ter-minal epitope is somehow masked during fixation, as we pre-viously reported [9].The immunohistochemical staining was performed using anAutomated Secondary System (cat #760-4205; Ventana,Tucson, AZ, USA). Slides were stained concurrently for ER,HER-2, and progesterone receptor (PR) using standard immu-noperoxidase techniques, as described previously [5,17].Breast cancer subtypes were determined using immunohisto-chemical markers (HER-2, ER, and PR) to define each type.Subtype definitions were as follows: luminal A (ER-positiveand/or PR positive, HER-2 negative), luminal B (ER positiveand/or PR positive, HER-2 positive), triple negative (ER nega-Page 2 of 9(page number not for citation purposes)nomas within 52 months [14]. tive, PR negative, HER-2 negative) [18], HER-2 positive (ERAvailable online http://breast-cancer-research.com/content/10/5/R86negative, PR negative, HER-2 positive), and unassigned (miss-ing data on any of the three markers). TMAs were evaluatedindependently by two pathologists (Drs Hamid Masoudi andDimithry Turbin) to quantify the percentage of tumor cellsstaining positive for YB-1, ER, and PR as follows: negative(<1%), positive 1+ (1% to 25%), positive 2+ (25% to 75%),or positive 3+ (>75%). For the analyses, YB-1 expression wasdichotomized into 0 or 1 (where 0 = negative or very low, and1 = moderate to high [2+ and 3+]). HER-2 was consideredpositive if it was 3+, as we previously described [9]. Of the4,049 samples, for 3,097 YB-1 could be scored and breastcancer specific survival (BCSS) outcome data were availablefor these. In addition, in 1,201 cases that could be scored forYB-1, data were available regarding relapse free survival(RFS). Immunohistochemical data was collected for samplesthat were blinded as to clinical outcomes.Statistical analyses were performed using SPSS software ver-sion 13.0 (SPSS Inc., Chicago, IL, USA) and were describedpreviously [5]. Briefly, for univariate analyses, BCSS and RFSwere estimated using Kaplan-Meier curves. Differences in sur-vival were determined using Breslow tests. For BCSS, survivaltime was censored at death if the cause was not breast canceror if the patient was still alive at the end of the study. Sixpatients with unknown cause of death were excluded from theBCSS analysis. In addition, RFS was also censored at death ifthe cause was not breast cancer and if the patients were alivewithout any relapse at the end the study. For multivariate anal-yses, Cox proportional hazards models were used to calculateadjusted hazard ratios (HRs), accounting for independent cov-ariates. A likelihood ratio was used to determine the extent ofassociation of YB-1 expression with particular subtypes.Spearman's correlation was also used to determine the extentof correlation between YB-1 expression and other markerssuch as ER.Results and discussionYB-1 was detected by immunostaining in 41% (1,644/4,049)of the patient samples, in which it was highly predictive ofdecreased RFS (P < 2.5 × 10-20). We observed that 80% ofpatients who expressed little or no YB-1 were relapse freeover 5 years, as compared to 60% if the protein was highlyexpressed (Figure 1a). Consistent with this observation, YB-1was associated with poor overall BCSS (P < 7.3 × 10-26; Fig-ure 1b). Among patents expressing low levels of YB-1, about90% were alive after 5 years, as compared with only 75% ofthose who expressed YB-1 (Figure 1b). Furthermore, this wasthe case whether patients were node positive or node nega-tive (Additional file 1). Specifically, YB-1 was expressed in33% of node-negative breast cancer patients, in whom it wasassociated with poor survival (P = 1.0 × 10-13). It was similarlyexpressed in 37% of node-positive patients, in whom it onceagain predicted poor survival (P = 3.6 × 10-14).Of note, patients whose tumors expressed YB-1 tended to beyounger (r = -0.203, P = 6.512 × 10-17), their tumors were ofhigher grade (r = 0.268, P = 2.957 × 10-27), and many wereER negative (r = -0.343, P = 5.452 × 10-47; Additional file 2).There was also a positive correlation with tumors harboringamplified HER-2 (r = 0.217, P = 8.588 × 10-19; Additional file2) corroborating smaller reports by us [9] and others [19], inwhich cohorts of 490 and 73 patients were evaluated,respectively.Importantly, YB-1 staining was consistently associated withdisease recurrence, independent of breast cancer subtype asdefined by ER, PR, and HER-2 expression (Figure 2: luminal A,P = 3.04 × 10-8; luminal B, P = 0.133; triple negative, P = 1.9× 10-2; HER-2, P = 2.7 × 10-2). An exception was found in theluminal B cohort, and the lack of significance was possiblybecause there were only 86 patients in this subgroup and sothe analysis was under-powered. Taking our observations fur-ther, BCSS was shorter when YB-1 was higher in all breastcancer subtypes (Figure 3: luminal A, P = 2.8 × 10-9; luminalB, P = 0.034; triple negative, P = 7 × 10-3; HER-2, P = 1.6 ×10-2). Of note is the observation that although YB-1 isexpressed less frequently in the luminal A subtype, its expres-sion was the most significant, probably because this consti-tuted the largest group of patients overall. Whether YB-1cooperates with estrogen to promote the growth of luminal Abreast cancer is not known and is a question that we areactively addressing. To evaluate further the significance of YB-1 as a prognostic factor across breast cancer subtypes, weconducted Cox regression analyses and confirmed that it inde-pendently predicted RFS (HR = 1.284, 95% confidence inter-val [CI] = 1.068 to 1.544; P = 8 × 10-3) and BCSS (HR =1.452, 95% CI = 1.253 to 1.682; P = 6.74 × 10-7; Additionalfiles 3 and 4).These findings prompted us to examine how YB-1 comparedwith established prognostic factors used to guide patient diag-nosis and thus treatment. Importantly, YB-1 was more indica-tive of RFS (HR = 1.331, 95% CI = 1.169–1.516; P = 1.61× 10-5) than HER-2 (HR = 1.256, 95% CI = 1.066 to 1.479;P = 6.0 × 10-3) or ER (HR = 0.942, 95% CI = 0.816 to 1.087;P = 4.13; Table 1). Similarly, YB-1 expression posed an evengreater risk for reduced BCSS (HR = 1.456, 95% CI = 1.257to 1.656; P = 5.6 × 10-7) than did HER-2 expression (HR =1.259, 95% CI = 1.052 to 1.506; P = 1.2 × 10-2) or ER status(HR = 0.816, 95% CI = 0.720 to 0.991; P = 0.038; Table 2).Moreover, YB-1 was also better at predicting outcome thantumor grade (HR = 1.416, 95% CI = 1.215 to 1.651; P = 8.9× 10-6) or patient age (HR = 1.160, 95% CI = 1.000 to 1.345;P = 5 × 10-2). We therefore conclude that YB-1 expressioncan be used to identify those patients who are likely to requiremore aggressive therapy, given its pronounced associationwith markedly reduced RFS and attenuated BCSS.Page 3 of 9(page number not for citation purposes)Breast Cancer Research    Vol 10 No 5    Habibi et al.Figure 1YB-1 expression is associated with relapse and poor survival in the entire cohort. (a) Out of a total of 1,201 patients, YB-1 was expressed in 41% (495/1201) with recurrent breast cancer. YB-1 expression was associated with shorter relapse free survival. 'N' represents the number of patients who had breast cancer recurrence (number of events). (b) The expression of YB-1 is associated with a significant increase in breast cancer specific deaths (BCSS). YB-1 was expressed in 35% (1,081/3,097) of patients with reduced BCSS. Patient survival was based on deaths specifically related to breast cancer and not other causes. BCSS = breast cancer specific survival; YB-1 = Y-box binding protein-1.Figure 2YB-1 is linked to relapse in all subtypes of breast cancer defined by hormone receptors and HER-2 statusl  r l  i  al  subtypes of breast cancer defined by hormone receptors and HER-2 status. Tumor tissue microarrays were stained for YB-1, ER, PR, and HER-2 (n = 2,978 cases in total for all markers). The relative distribution of YB-1 was evaluated in breast cancer subtypes. In all subtypes, the expression of YB-1 was associated with shorter relapse free survival. 'N' represents the number of YB-1 positive patients who had breast cancer relapse (number of events) in each subtype. ER, estrogen receptor; HER, human epidermal growth factor receptor; PR, progesterone Page 4 of 9(page number not for citation purposes)receptor; YB-1, Y-box binding protein-1.Available online http://breast-cancer-research.com/content/10/5/R86As breast cancer is being diagnosed earlier, there is a need toidentify those individuals who may or may not need chemother-apy. Patients who have grade I to II tumors with no evidencefor lymph node involvement (T1N0) have been categorized asbeing in a low-risk group, in which the odds of survival shouldbe very good. Within the TMA we analyzed, there were 1,292cases considered to be low risk. However, if these patientshad tumors that expressed YB-1 they were more likely to diefrom breast cancer (Figure 4). We found that 437 out of 1,292(34%) cases expressed YB-1, indicating that a significant pro-portion of patients considered to be at low risk were morelikely to die from the disease than expected, suggesting thatmore aggressive treatment might have improved their out-come. Using 5-year survival as a benchmark, we found thatmore than 95% of patients were alive if YB-1 was not highly75% of patients were alive at 5 years (Figure 4; P = 7.09 × 10-17). By multivariate analysis, YB-1 expression was a significantadverse prognostic biomarker in patients who did not receivechemotherapy (HR = 1.898, 95% CI = 1.460 to 2.465; P <1.67 × 10-6; Additional file 5). Again, it was better at predictingBCSS than HER-2 or ER (Additional file 5).To further this line of investigation, we addressed whether YB-1 emerged as a prognostic marker in patients taking tamoxifen.In a nested cohort of ER-positive breast cancer patientstreated with tamoxifen (n = 671), YB-1 expression was asso-ciated with reduced BCSS (Figure 5). Of note, YB-1 wasexpressed in 185 out of 671 cases (28%), indicating thatexpression of this protein is not an uncommon event in thesetumors. We chose to evaluate this within a 5-year time-frame,Figure 3YB-1 is associated with BCSS in all subtypes of breast cancerit   i  al  subtypes of breast cancer. YB-1 is expressed in 23.1% (480/2078) of luminal A and 51.6% (95/183) of luminal B, and is more likely to be expressed in the ER-negative breast cancers, in which it was found in 64.6% (320/493) and 66.4% (148/223) of TNP and HER-2 subtypes, respectively. Its expression is associated with poor survival in all subtypes of breast cancer. 'N' represent of number of YB-1 positive patients who died from breast cancer (number of events) in each subtype. Examples of the immunostaining are represented in the inset (magnification 200×). BCSS, breast cancer specific survival; ER, estrogen receptor; YB-1, Y-box binding protein-1.Page 5 of 9(page number not for citation purposes)expressed. However, if YB-1 expression was high then only given that this is the typical duration of tamoxifen administra-Breast Cancer Research    Vol 10 No 5    Habibi et al.tion. In a Cox regression model, YB-1 (HR = 1.707; P = 2.2 ×10-2) was independently associated with a greater risk forbreast cancer related death (Additional file 6). Thus, in theadjuvant setting, YB-1 can also serve as a useful biomarker ofaggressive disease. It would therefore be interesting to evalu-ate YB-1 within the context of more contemporary standard-of-care hormone therapies such as raloxifene, letrozole, anastro-zole, and exemestane.Our data indicate that YB-1 is a strong prognostic marker forall subtypes of human breast cancer, even in cases that werethought to be low risk and therefore did not need chemother-apy. We also extend preliminary clinical data from others sug-gesting that YB-1 expression distinguish a high-risk group ofbreast cancer patients. Notably, Janz and coworkers [20]observed YB-1 expression in 49% (20/42) of patients receiv-ing chemotherapy, and there was a trend toward an associa-tion of YB-1 with poor disease-free survival, although thisfinding did not reach statistical significance. The same authorsalso examined YB-1 expression in a cohort of 42 low-risk can-the patients who had low YB-1 expression relapsed, whereas30% of those with high levels did (P < 1.1 × 10-2). In addition,Huang and colleagues [21], in a study of 42 patients, reportedthat YB-1 was associated with recurrence after adjuvantchemotherapy, although long-term outcomes were not evalu-ated. We now provide definitive retrospective data that indi-cate that YB-1 expression in low-risk patients is stronglyassociated with a higher rate of breast cancer related deaths.Importantly, YB-1 was a significant factor in a Cox regressionmodel when other variables that typically define risk weretaken into consideration. Taking these analyses together withthe prognostic value of YB-1 identified in the present study,YB-1 was comparable to tumor size and was superior to HER-2, ER, patient age, or tumor grade as a prognostic marker forRFS and BCSS in the majority of analyses.During the past few years, several new approaches have beenproposed to identify patients at high risk. For example, diag-nostic tests have been developed to stratify patients based ongene expression and are currently being prospectively evalu-Table 1In a multivariate analysis of recurrent breast cancer cases, the expression of YB-1 was better than HER-2 or ER in predicting eventsPrognostic marker Patients with relapseHR (95% CI) P valueNodal statusPositive versus negative 1.951 (1.725 to 2.206) 1.7 × 10-26Tumor size (cm)2 to 5 1.392 (1.227 to 1.579) 2.8 × 10-7>5 1.921 (1.227 to 2.445) 1.2 × 10-7GradeIII versus II versus I 1.300 (1.139 to 1.482) 9.6 × 10-5Age (years)<50 versus ≥ 50 1.000 (0.878 to 1.139) 1.000ERPositive versus negative 0.942 (0.816 to 1.087) 0.41HER-2Positive versus negative 1.256 (1.066 to 1.479) 6.0 × 10-3YB-1Positive versus negative 1.331 (1.169 to 1.516) 1.6 × 10-5The predictive power of YB-1 was similar to that of tumor grade and smaller tumor size, but not as great as that of nodal status. CI, confidence interval; ER, estrogen receptor; HER, human epidermal growth factor receptor; HR, hazard ratio; YB-1 = Y-box binding protein-1.Table 2In a multivariate analysis of breast cancer cases, the expression of YB-1 was better than HER-2 or ER in predicting reduced BCSSPrognostic marker Patients with reduced BCSSHR (95% CI) P valueNodal statusPositive versus negative 2.405 (2.083 to 2.776) 4.6 × 10-33Tumor size (cm)2 to 5 1.605 (1.387 to 1.859) 2.8 × 10-10>5 2.269 (1.749 to 2.943) 2.8 × 10-10GradeIII versus II versus I 1.416 (1.215 to 1.651) 8.9 × 10-6Age (years)<50 versus = 50 1.160 (1.000 to 1.345) 5.0 × 10-2ERPositive versus negative 0.816 (0.720 to 0.991) 3.8 × 10-2Her-2Positive versus negative 1.259 (1.052 to 1.506) 1.2 × 10-2YB-1Positive versus negative 1.456 (1.257 to 1.656) 5.6 × 10-7YB-1 was approximately equal to the power of tumor grade yet not as accurate as tumor size or nodal status in predicting reduced BCSS. BCSS, breast cancer specific survival; CI, confidence interval; ER, estrogen receptor; HER, human epidermal growth factor receptor; HR, hazard ratio; YB-1 = Y-box binding protein-1.Page 6 of 9(page number not for citation purposes)cers not treated with chemotherapy and found that it wasexpressed in 76% (32/42) of those cases. Notably, none ofated. The first test, Mammaprint®, stratifies node-negativebreast cancer patients based on a set of 70 genes usingAvailable online http://breast-cancer-research.com/content/10/5/R86cDNA microarrays [22,23]. A quantitative PCR-basedmethod, Oncotype Dx®, is commercially available and appearsto define a group of low-risk patients who may be able to avoidchemotherapy [24]. This test was designed to amplify a cas-sette of 21 genes (16 target genes and five housekeepinggenes). Interestingly, within this cassette there are two knownYB-1 target genes, namely cyclin B and HER-2. Furthermore,the assay is designed to detect Ki67, which is associated withhigher levels of YB-1 [9].Because YB-1 is expressed in 30% of node-negative breastcancers, based on the data presented herein, it may be usefulin stratifying patients thought to be at low risk. We proposethat YB-1 could be monitored either using quantitative RT-PCR or in a low-cost manner using immunohistochemistry, justas HER-2 and ER proteins are currently evaluated clinically.We have already optimized the detection of YB-1 using quan-titative RT-PCR, as described previously [9], and by immunos-taining using an automated system described herein. Thus,this novel biomarker could easily be translated into routine clin-ical practice for diagnostic purposes. Perhaps YB-1 may serveto define further breast cancer subtypes in general, becausewe are now beginning to appreciate that triple negative breastcancers can be further divided into five subcategories [18].Thus, YB-1 can be added to define high-risk patients withinthe triple negative subtype, as was recently reported for EGFR[25]. Finally, the degree to which the nuclear localization ofYB-1 contributes to the aggressiveness of breast cancerscontinues to be of great interest to us, given its establishedrole as an oncogenic transcription factor. Additional studieswill be required, with antibodies that detect nuclear YB-1, toshed light on this.Although we found that YB-1 is expressed in about 40% ofinvasive breast carcinomas, the underlying reason for this isunknown. Evidence thus far indicates that YB-1 is not com-monly amplified, based on analyses of primary tumors [16] andbreast cancer cell lines [26]. There are reports indicating thatGATA-1 binds to the YB-1 promoter and suppresses itsexpression during erythrocyte differentiation [27]. Understress induced by cisplatin c-myc can also bind to E-boxeslocated at several sites along the YB-1 promoter [28].Although stress can induce YB-1, it is still curious to us thatbreast cancer cell lines express high levels under normalgrowth conditions, and patients have high levels before chem-otherapy or radiation because the tumors from that were underinvestigation in this report were obtained prior to treatment. Tobegin to elucidate the regulation of YB-1, we examined its pro-moter for potential regulatory elements using CONSITE (a pre-dictive algorithm that mitigates against false positives) [29].Using a 100% stringency criterion, we found that across spe-cies the YB-1 promoter has n-myc and snail binding sites [29].Furthermore, we identified additional regulatory sites on theYB-1 promoter, such as hunchback, rel, sox, and myf [29].Although n-myc is not thought to play a central role in breastcancer, c-myc does play a role [30], and they both bind to E-boxes. Our virtual screen is consistent with the initial charac-terization of the YB-1 promoter, in which six E-boxes wereidentified [31]; this may explain why c-myc reportedly inducesits expression in cooperation with p73 [28]. More recently,twist – which also binds to E-boxes – was reported to induceFigure 4YB-1 is a high risk factor in low-risk patientsi ri  f t r i  lo -risk patients. YB-1 is expressed in 34% (437/1292) of low-risk patients who received no systemic therapy. 'N' represents the number of patients who died from breast cancer (number of events) in each category. Survival was stratified based on whether tumors expressed YB-1 (light grey line) or did not (black line). YB-1, Y-box binding protein-1.Figure 5In the adjuvant setting expression of YB-1 is indicative of reduced BCSSBCSS. YB-1 expression was evaluated in a nested cohort of 671 tamoxifen-treated patients. Twenty-eight percent of patients expressed YB-1 (185/671), in whom its expression was associated with reduced BCSS. Survival was assessed over 5 years, because this is the stand-ard duration of tamoxifen administration. ER, estrogen receptor; YB-1, Y-box binding protein-1.Page 7 of 9(page number not for citation purposes)Breast Cancer Research    Vol 10 No 5    Habibi et al.YB-1 [32]. To date, this is the only study addressing the regu-lation of YB-1 expression specifically in breast cancer.ConclusionsThe present data provide new impetus to translate these labo-ratory findings into routine clinical practice, given the remarka-ble strength of YB-1 as a biomarker for aggressive breastcancer across all subtypes. We anticipate that this newbiomarker will create better opportunities to individualize ther-apy and may ultimately be developed as a novel moleculartarget.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsSED and GH were responsible for study design, data analysis,and manuscript preparation. DT, HM, TN, and DH wereresponsible for pathology evaluations. GH, SL, JHL, KG, andMP were responsible for statistical analyses. DH, KG, TN, HM,and DT were responsible for TMA construction and clinicalcorrelates.Additional filesAcknowledgementsFunding was provided by the Canadian Breast Cancer Research Alli-ance through two grants the National Cancer Institute of Canada and the Translational Acceleration Grant-II (TAG-II) awarded to SED. Addi-tional grant support was provided through the TAG-I to KAG and DGH. We would like to thank Dr C Blake Gilks for his continued support on this project and for his mentorship to Ms Golareh Habibi.References1. Sawyers C: The cancer biomarker problem.  Nature 2008,452:548-552.2. Harris L, Fritsche H, Mennel R, Norton L, Ravdin P, Taube S,Somerfield MR, Hayes DF, Bast RCJ: American Society for Clin-ical Oncology 2007: update of recommendations for the use oftumor markers in breast cancer.  J Clin Oncol 2007,The following Additional files are available online:Additional file 1This file shows that YB-1 is strongly associated with poor survival in node-positive and node-negative breast cancers. In node-negative tumors YB-1 was expressed in 33% (572/1,730) of the cases, in which it was strongly associated with reduced BCSS. Similarly, it was expressed in 37% (506/1356) of node-positive breast cancer cases. In these cases YB-1 was also positively associated with reduced BCSS.See http://www.biomedcentral.com/content/supplementary/bcr2156-S1.jpegAdditional file 2This file shows correlations between YB-1 and clinicopathological features of breast cancer. Patients who had tumors expressing YB-1 were younger and tended to have tumors that lacked the ER. There was a positive correlation with increasing tumor grade. Often, tumors that exhibited amplifications in HER-2 also expressed high levels of YB-1. Conversely, YB-1 was not associated with nodal status and weakly related to tumor size.See http://www.biomedcentral.com/content/supplementary/bcr2156-S2.jpegAdditional file 3This file shows that YB-1 is significantly associated with relapse, independent of the type of breast cancer. The expression of YB-1 was associated with shorter RFS (HR = 1.284; P = 0.008), independent of breast cancer subtype defined by hormone receptor and HER-2 status, based on a Cox regression analysis.See http://www.biomedcentral.com/content/supplementary/bcr2156-S3.jpegAdditional file 4This file shows that the prognostic significance of YB-1 is associated with reduced BCSS (HR = 1.46, P = 6.74 × 10-7), independent of tumor subtype.See http://www.biomedcentral.com/content/supplementary/bcr2156-S4.jpegAdditional file 5This file shows a Cox regression model for patients who were treated with surgical resection and no chemotherapy. Nodal status, tumor size, and YB-1 expression were associated with reduced BCSS. YB-1 was better than HER-2 or ER for predicting BCSS.See http://www.biomedcentral.com/content/supplementary/bcr2156-S5.jpegAdditional file 6This file shows a Cox regression analysis for ER-positive patients treated with tamoxifen for 5 years. YB-1 was independently associated with an increased risk for reduced BCSS (HR = 1.703, P = 0.022). YB-1 complemented the significance found in node status, tumor size (greater than 2 cm only), and grade. Patient age, small tumors (<2 cm), and HER-2 expression were not independently associated with reduced BCSS.See http://www.biomedcentral.com/content/supplementary/bcr2156-S6.jpegPage 8 of 9(page number not for citation purposes)25:5287-5312.3. Viale G, Regan MM, Mastropasqua MG, maffinin F, Maiorano E,Colleoni M, Price KN, Golouh R, Perin T, Brown RW, Kovacs A, Pil-Available online http://breast-cancer-research.com/content/10/5/R86lay K, Ohsschlegel C, Gusterston BA, Castiglione-Gertsch M, Gel-ber RD, Goldhirsch A, Coates AS: Predictive value of tumor Ki-67 expression in two randomized trials of adjuvant chemoen-docrine therapy for node-negative breast cancer.  J Natl Can-cer Inst 2008, 100:207-212.4. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuireWL: Human breast cancer: correlation of relapse and survivalwith amplification of HER-2/neu oncogene.  Science 1987,235:177-182.5. Cheang MCU, Treaba DO, Speers CH, Olivotto IA, Bajdik CD,Chia SK, Goldstein LC, Gelmon KA, Huntsman DG, Gilks CB,Nielsen TO, Gown AM: Immunohistochemical detection usingthe new rabbit monoclonal antibody SP1 of estrogen receptorin breast cancer is superior to mouse monoclonal 1D5 in pre-dicting survival.  J Clin Oncol 2006, 24:5637-5644.6. McGuire WL: Breast cancer prognostic factors: evaluationguidelines.  J Natl Cancer Inst 1991, 83:154-155.7. Ellis MJ: Impoving outcomes for hormone receptor-positivebreast cancer: back to the drawing board.  J Natl Cancer Inst2008, 100:159-161.8. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnson H,Hastie T, Eisen M, Rijn M van de, Jeffrey SS, Thorsen T, Quist H,Matase JC, Brown PO, Botstein D, Lonning PE, Borresen-Dale A:Gene expression patterns of breast carcinomas distinguishtumor subclasses with clinical implications.  Proc Nat Acad SciUSA 2001, 98:10869-10874.9. Wu J, Lee C, Yokom D, Jiang H, Cheang MCU, Yorida E, Turbin D,Berquin IM, Mertens PR, Iftner T, Gilks B, Dunn SE: Disruption ofthe Y-box binding protein-1 (YB-1) results in suppression ofthe epidermal growth factor receptor and Her-2.  Cancer Res2006, 66:4872-4879.10. Ise T, Nagatani G, Imamura T, Kato K, Takano H, Naomoto M, IzumiH, Ohmori H, Okamoto T, Ohga T, Uchiumi T, Kuwano M, KohnoK: Transcription factor Y-box binding protein 1 binds preferen-tially to cisplatin-modified DNA and interacts with proliferatingcell nuclear antigen.  Cancer Res 1999, 59:342-346.11. Bargou RC, Jurchott K, Wagener C, Bergmann S, Metzner S,Bommert K, Mapara MY, Winzer KJ, Dietel M, Dorken B, Royer HD:Nuclear localization and increased levels of transcription fac-tor YB-1 in primary human breast cancers are associated withintrinsic MDR1 gene expression.  Nat Med 1997, 3:447-450.12. Jurchott K, Bergmann S, Stein U, Walther W, Janz M, Manni I,Paiggo G, Fietze E, Dietel M, Royer HD: Yb-1 as a cell cycle reg-ulated transcription factor facilitating cyclin A and cyclin B1gene expression.  J Biol Chem 2003, 278:27988-27996.13. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, DengS, Johnsen H, Pesich R, Geisler S, Demeter J, Perou CM, LonningPE, Brown PO, Borresen-Dale A, Botstein D: Repeated observa-tion of breast tumor subtypes in independent gene expressiondata sets.  Proc Nat Acad Sci USA 2003, 100:8418-8423.14. Bergmann S, Royer-Pokora B, Fietze E, Jürchott K, Hildebrandt B,Trost D, Leenders F, Claude J, Theuring F, Bargou R, Dietel M,Royer HD: YB-1 provokes breast cancer through the inductionof chromosomal instability that emerges from mitotic failureand centrosome amplification.  Cancer Res 2005,65:4078-4087.15. Sutherland BW, Kucab JE, Wu J, Lee C, Cheang MCU, Yorida E,Turbin D, Dedhar S, Nelson CC, Pollack M, Grimes HL, Miller K,Badve S, Huntsman D, Gilks B, Chen M, Pallen CJ, Dunn SE: Aktphosphorylates the Y-box binding protein-1 at Ser102 locatedin the cold shock domain and affects the anchorage-inde-pendent growth of breast cancer cells.  Oncogene 2005,24:4281-4292.16. Stratford AL, Habibi G, Astanehe A, Jiang H, Hu K, Park E, ShadeoA, Buys TPH, Lam WL, Pugh T, Marra M, Nielsen TO, Klinge U,Mertens PR, Aparicio S, Dunn SE: Epidermal growth factorreceptor (EGFR) is transcriptionally induced by the Y-boxbinding protein-1 (YB-1) and can be inhibited with Iressa inbasal-like breast cancer providing a potential target fortherapy.  Breast Cancer Res 2007, 9:R61.17. Rajput AB, Turbin DA, Cheang MCU, Vodue DK, Leung S, GelmonKA, Gilks CB, Huntsman DG: Stromal mast cells in invasivebreast cancer are a marker of favorable prognosis: A study of4,444 cases.  Breast Cancer Res Treat 2007, 107:249-57.topathological characterization of triple-negative/basal-likebreast carcinomas.  Breast Cancer Res 2007, 9:R65.19. Fujii T, Kawahara A, Basaki Y, Hattori S, Nakashima K, Nakano K,Shirouzu K, Kohno K, Yanagawa T, Yamana H, Nishio K, Ono M,Kuwano M, Kage M: Expression of Her-2 and estrogen receptoralpha depends upon nuclear localization of Y-box binding pro-tein-1 in human breast cancers.  Cancer Res 2008,68:1504-1512.20. Janz M, Harbeck N, Dettmar P, Berger U, Schmidt A, Jurchott K,Schmitt M, Royer HD: Y-box factor YB-1 predicts drug resist-ance and patient outcome in breast cancer independent ofclinically relevant tumor biologic factors HER2, uPA and PAI-1.Int J Cancer 2002, 97:278-282.21. Huang J, Tan PH, Li KB, Matsumoto K, Tsujimoto M, Bay BH: Y-box binding protein, YB-1, as a marker of tumor aggressive-ness and response to adjuvant chemotherapy in breastcancer.  Int J Oncol 2005, 26:607-613.22. Vijver JM Van de, He YD, Van't Veer LJ, Dai H, Hart AA, Voskill DW,Schreiber GJ, Peterse JL, Roberts CT, Marton MF, Parrish M,Atsma D, Witteveen AT, Glas AM, Delahaye LJMJ, Velde T Van der,H B, Rodenhauis S, Rutgers ET, Friend SH, Bernards R: A geneexpression signature as a predictor of survival in breastcancer.  N Engl J Med 2002, 347:1999-2009.23. Glas AM, Floore A, Delahaye LJMJ, Witteveen AT, Pover RCF,Bakz N, Lahti-Domenici JST, Bruinsma TJ, Warmoes MO, BernardsR, Wessels LFA, Van 't Veer LJ: Converting a breast cancermicroarray signature into a high-throughput diagnostic test.BMC Genomics 2006, 7:278.24. Paik S, Shak S, Tang G, Kim G, Baker J, Cronin M, Baehner F,Walker MG, Watson D, Park T, Hiller W, Risher WR, WickermanD, Bryant J, Wolmark N: A multigene assay to predict recur-rence of tamoxifen-treated, node-negative breast cancer.  NEngl J Med 2004, 351:2817-2826.25. Cheang MCU, Voduc D, Bajdik C, Leung S, McKinney S, Chia SK,Perou CM, Nielsen TO: Basal-like breast cancer defined by fivebiomarkers has superior prognostic value than triple-negativephenotype.  Clin Cancer Res 2008, 14:1368-1376.26. Shadeo A, Lam WL: Comprehensive copy number profiles ofbreast cancer model genomes.  Breast Cancer Res 2006, 8:R9.27. Yokoyama H, Harigae H, Takahashi S, Takahashi S, Furyuama K,Kaku M, Yamamoto M, Sasaki T: Regulation of YB-1 geneexpression by GATA transcription factors.  Biochem BiophyRes Commun 2003, 303:140-145.28. Uramoto H, Izumi H, Ise T, Tada M, Uchiumi T, Kuwano M, Yas-umoto K, Funa K, Kohno K: p73 interacts with c-myc to regulateY-box binding protein-1 expression.  J Biol Chem 2002,277:31694-31702.29. Wu J, Stratford AL, Astanehe A, Dunn SE: YB-1 is a transcrip-tion/translation factor that orchestrates the oncogenome byhardwiring signal transduction to gene expression.  Transla-tional Oncogenomics 2007, 2:1-17.30. Efstratiadis A, Szabolcs M, Klinakis A: Notch, myc and breastcancer.  Cell Cycle 2007, 6:418-429.31. Makino Y, Ohga T, Toh s, Koike K, Okumura K, Wada M, KuwanoM, Kohno K: Structural and functional analysis of the human Y-box binding protein (YB-1) gene promoter.  Nucleic Acid Res1996, 24:1873-1878.32. Shiota M, Izumi H, Onitshka T, Miyamoto N, Kashiwagi E, Kidani A,Yokomizo A, Naito S, Kohno K: Twist promotes tumor cellgrowth through YB-1 expression.  Cancer Res 2008,68:98-105.Page 9 of 9(page number not for citation purposes)18. Kreike B, van Kouwenhove M, Horlings H, Weigelt B, Peterse H,Bartelink H, Vijver MJ van de: Gene expression profiling and his-


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items