|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 6 (February 20), 2007: pp. 662-668 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.07.3676 HOXB13-to-IL17BR Expression Ratio Is Related With Tumor Aggressiveness and Response to Tamoxifen of Recurrent Breast Cancer: A Retrospective Study
From the Department of Medical Oncology, Erasmus MC/Daniel den Hoed Cancer Center, Rotterdam, the Netherlands Address reprint requests to Maurice P.H.M. Jansen, PhD, Erasmus MC, Department of Medical Oncology, Josephine Nefkens Institute, PO Box 2040, 3000 CA Rotterdam, the Netherlands; e-mail: m.p.h.m.jansen{at}erasmusmc.nl
PURPOSE: A HOXB13-to-IL17BR expression ratio was previously identified to predict clinical outcome of breast cancer patients treated with adjuvant tamoxifen. However, this ratio may predict a tumor's response to tamoxifen, its intrinsic aggressiveness, or both. PATIENTS AND METHODS: We have measured the HOXB13 and IL17BR expression levels by real-time polymerase chain reaction in 1,252 primary breast tumor specimens. Expression levels were normalized to housekeeper gene levels and related to clinicopathologic factors for all patients. The primary objective of this study was to determine the relationship of a HOXB13-to-IL17BR ratio with tumor aggressiveness and/or with response to tamoxifen therapy in estrogen receptor (ER) -positive disease. We selected ER-positive tumors, and clinical end points for the HOXB13-to-IL17BR ratio were disease-free survival (DFS) in patients with primary breast cancer (N = 619) and progression-free survival (PFS) in patients with recurrent breast cancer treated with first-line tamoxifen monotherapy (N = 193). The odds ratio (OR) and hazard ratio (HR) and their 95% CI were calculated, and all P values were two-sided. RESULTS: The HOXB13-to-IL17BR ratio was significantly associated with DFS and PFS. In multivariate analysis, HOXB13-to-IL17BR ratio expression levels were associated with a shorter DFS for node-negative patients only. Corrected for traditional predictive factors, the dichotomized HOXB13-to-IL17BR ratio was the strongest predictor in multivariate analysis for a poor response to tamoxifen therapy (OR = 0.16; 95% CI, 0.06 to 0.45; P < .001) and a shorter PFS (HR = 2.97; 95% CI, 1.82 to 4.86; P < .001). CONCLUSION: High HOXB13-to-IL17BR ratio expression levels associate with both tumor aggressiveness and tamoxifen therapy failure.
Approximately 70% to 75% of breast tumors express the estrogen receptor (ER), the classical prognostic factor, which is an important target for endocrine therapy. In the adjuvant setting, tamoxifen therapy results in a 5.3% to 12.6% improvement in 10-year survival in lymph node-negative (LNN) and lymph node-positive (LNP) patients, respectively.1 In recurrent disease, approximately half of the patients with ER-positive primary breast tumors will not respond or will rapidly develop resistance to tamoxifen. Based on genome-wide screening, signatures associated with response to tamoxifen therapy of ER-positive breast cancer have been published by Ma et al2 for adjuvant treatment and published by Jansen et al3 for recurrent disease. Ma et al2 identified a two-gene expression ratio, HOXB13-to-IL17BR, which predicted clinical outcome in a retrospective study of 60 patients and confirmed this in formalin-fixed paraffin-embedded (FFPE) samples of 20 LNN patients. Reid et al,4 however, failed to validate this two-gene ratio on frozen samples from 58 patients, who were mainly node-positive. A drawback in the adjuvant setting is that the tumor's response to tamoxifen and its intrinsic aggressiveness are measured. No accurate determination of response to tamoxifen can be given because these studies lacked randomization and a nontreated control group. Recent microarray data analyses showed that the HOXB13-to-IL17BR ratio has no association with relapse-free survival5 but it does have a weak association with response to first-line tamoxifen therapy.6 However, data from microarray experiments need to be confirmed with quantitative real-time polymerase chain reaction (qRT-PCR), as these data are considered to be more accurate.4 Therefore, we measured mRNA expression levels of HOXB13 and IL17BR with a qRT-PCR in RNA isolated from a set of 1,252 frozen breast cancer specimens and related expression levels with clinicopathologic factors. The main clinical end points in ER-positive disease were (1) disease-free survival (DFS) of untreated LNN patients to determine tumor aggressiveness and (2) response to first-line tamoxifen monotherapy for recurrent disease. In addition, we compared for IL17BR two primer sets (ps): one at the 3' end region, comparable to the assay of Ma et al (ps3),2 and one at the 5' end region, used by Reid et al (ps5).4 Finally, we validated in our cohort predefined cutoff points for untreated patients and tamoxifen-treated patients.7,8
Patients This retrospective study was approved by the medical ethics committee of the Erasmus MC (Rotterdam, the Netherlands; MEC 02.953), and it included breast tumor tissue specimens of 1,683 female patients with primary operable breast cancer. To avoid bias, frozen tumor samples were processed from patients with breast cancer who entered the clinic between 1979 and 1996 and from whom detailed clinical follow-up was available. Follow-up, tumor staging, and response to therapy was defined by standard International Union Against Cancer (Geneva, Switzerland) classification criteria9 and applied previously by Foekens et al.10 The cutoff point to classify primary breast tumors as ER and/or progesterone receptor (PgR) -positive was 10 fmol/mg cytosolic protein. The following tumors were excluded from analysis: (1) with distant spread at or within the first month of surgery; (2) with missing values for lymph node status, ER protein status, and HOXB13 and IL17BR expression; (3) with < 30% epithelial tumor cells in their tumor specimens; and (4) with specimens of poor RNA quality.11 After applying the exclusion criteria, tumor specimens of 1,252 patients (74%) were analyzed for HOXB13 and IL17BR expression. From these 1,252 patients (for clinicopathologic details, see Table 1), 543 patients (43%) underwent breast-conserving lumpectomy, and 709 patients (57%) underwent modified mastectomy. The median follow-up time of all 1,252 patients was 72 months (2 to 248 months), of the 692 patients alive was 91 months (3 to 248 months), and of the 560 deaths was 46 months (2 to 205 months). Disease recurrence occurred in 692 (55%) of 1,252 patients.
Four hundred six patients (32%) were treated with adjuvant systemic therapy, of which 177 patients (44%) received hormonal therapy, 214 (53%) received chemotherapy, and 15 (3%) received combination therapy. Of the 846 patients (68%) that had not received neoadjuvant systemic therapy, 381 patients (45%) experienced a relapse treated with systemic therapy. Two hundred eighty-five (75%) of these patients were treated with hormonal therapy, of which 193 patients (68%) received first-line tamoxifen monotherapy. Therapy failure occurred in 73 patients (38%), of which 61 patients (84%) had progressive disease and 12 patients (16%) showed stable disease  6 months. One hundred twenty patients (62%) of these 193 patients had experienced a clinical benefit from first-line tamoxifen monotherapy, of which 11 patients (66%) had a complete response, 30 patients (25%) showed a partial response, and 79 patients (66%) showed stable disease > 6 months.
RNA Isolation and qRT-PCR Forty rounds of amplification were performed according to the supplier's protocol, and at the end of the amplification fluorescent signals of the TaqMan probes (Applied Biosystems) were used to generate cycle threshold (Ct) values from which mRNA expression levels were calculated. Expression levels of HOXB13 and IL17BR were normalized against average expression levels of three reference genes as follows11: mRNA target = 2(mean Ct Ref –mean Ct Target). When HOXB13 amplification rounds exceeded the manufacturer's defined detection threshold of the ABI Prism 7700 Sequence Detection System (Ct values 35 to 40; Applied Biosystems, http://docs.appliedbiosystems.com/pebiodocs/04371095.pdf), quantities were considered to be undetectable and were set to 50% of the expression level measurable at the detection threshold.
Data Analysis and Statistics Only when the test for trend for a continuous variable was statistically significant, a cutoff point was considered justified. To define cutoff points, isotonic regression was used to find the points where the monotonic relationship between the measured level and the hazard ratio showed a distinct change. The cutoff with the largest statistically significant change in hazard ratio corrected for multiple testing was used.12 Ma et al7and Erlander et al8 have described cutoff points for untreated patients (cutoff = 1.00) and for adjuvant tamoxifen-treated patients (cutoff = 0.06) that separated (tamoxifen-treated) nonrecurrence cases from recurrent cases. These predefined cutoff points were not directly applicable to our normalized data set because of differences in assays, reference genes, qRT-PCR machines, and sample specimens. As a result, our data were standardized for each gene via a z-transformation step. In concordance with Ma et al,7 a composite index was generated by taking the difference between the standardized HOXB13 and IL17BR expression levels. Based on the cutoff points, survival curves were generated using the Kaplan-Meier method, and a log-rank test was used to test for differences. Computations were performed with the STATA statistical package, release 9.1 (STATA Corp, College Station, TX). All P values were two-sided, and P < .05 was considered statistically significant.
Associations of HOXB13 and IL17BR With Clinicopathologic Factors The mRNA expression levels of HOXB13, IL17BRps3, and IL17BRps5 were measured in 1,252 primary breast tumors by qRT-PCR and normalized against our reference genes. HOXB13 expression levels showed a weak but significant inverse association with those of IL17BR (Spearman's rho = –0.16; P < .0001), whereas expression levels of IL17BRps3 and IL17BRps5 correlated significantly (Spearman's rho = 0.91; P < .0001). In 448 tumors (36%), HOXB13 expression levels were below detection level (see Patients and Methods). Table 1 shows median expression levels and the interquartile range of all three transcripts and their relation with patient and tumor characteristics. The differences in IL17BR expression levels measured with ps3 and ps5 only reflect assay performance. HOXB13 levels associate significantly with grade and inversely with steroid hormone receptor status. The median expression level of HOXB13 was 17 x higher in poorly differentiated tumors compared with good/moderately differentiated tumors. In contrast, the median HOXB13 expression level was 25 x lower in ER-positive tumors compared with ER-negative tumors. Undetectable levels of HOXB13 were significantly more prevalent in ER-positive tumors, with 379 (41%) out of 917, compared with ER-negative tumors, with 69 (21%) out of 335; P < .001. Except for tumor size, IL17BR levels were significantly associated with all clinicopathologic parameters studied (ie, positively with age and menopausal status, and negatively with grade and nodal status). ER-positive tumors showed a two-fold higher median IL17BR expression level than ER-negative tumors. Next, expression levels of HOXB13 were divided by IL17BR to generate a HOXB13-to-IL17BR expression ratio. In all 1,252 tumors, the HOXB13-to-IL17BR ratio measured as univariate log-transformed continuous variable was associated with a poor DFS (HRps3 = 1.04 [95% CI, 1.02 to 1.06; P < .001] and HRps5 = 1.05 [95% CI, 1.03 to 1.06; P < .001], respectively) and a poor OS (HRps3 = 1.06 [95% CI, 1.04 to 1.08; P < .001] and HRps5 = 1.06 [95% CI, 1.04 to 1.08; P < .001], respectively).
HOXB13-to-IL17BR Ratio and Tumor Aggressiveness
The prognostic value of HOXB13-to-IL17BR ratios was also explored in ER-positive tumors from 151 LNP untreated patients, who were mainly enrolled in the early 1980s. In univariate analysis, the HOXB13-to-IL17BR ratio in these LNP patients associated with a poor DFS (HRps3 = 1.05 [95% CI, 1.01 to 1.09; P = .023], and HRps5 = 1.05 [95% CI, 1.01 to 1.10; P = .016], respectively) and a poor OS (HRps3 = 1.09 [95% CI, 1.04 to 1.14; P < .001] and HRps5 = 1.09 [95% CI, 1.04 to 1.14; P < .001], respectively). In the multivariate model, the HOXB13-to-IL17BR ratio was significantly associated with OS (P < .001), but not with DFS (P = .065). Based on a predefined cutoff point (1.00) for untreated patients7,8 a HOXB13-to-IL17BR index was dichotomized. In the LNN-untreated 468 patients cohort, this dichotomized index had a significant relationship with a poor DFS in univariate analysis (HRps3 = 1.69 and HRps5 = 1.59) and multivariate analysis (HRps3 = 1.74 and HRps5 = 1.61; Table 2). However, the dichotomized index was not related with DFS in the LNP-untreated cohort of 151 patients (data not shown).
HOXB13-to-IL17BR Expression Ratio and Response to First-Line Tamoxifen Monotherapy The HOXB13-to-IL17BR ratio, as a univariate continuous variable, was significantly related with a poor response (ORps3 = 0.93 [95% CI, 0.87 to 0.99]; P = .027 and ORps5 = 0.92 [95% CI, 0.86 to 0.98]; P = .015, respectively), a short PFS (Table 3), and a poor PRS (HRps3 = 1.07 [95% CI, 1.03 to 1.11; P < .001], and HRps5 = 1.07 [95% CI, 1.03 to 1.11; P < .001], respectively). In multivariate analysis, however, the HOXB13-to-IL17BR ratio retained only its significant association for PFS (Table 3) and PRS (P < .001; data not shown).
The significant findings in univariate analysis justified the search for a predictive cutoff point. Isotonic regression analysis defined optimal cutoff points for PFS at 2.99 and 16.44 for HOXB13-to-IL17BRps3 and IL17BRps5 ratio with 26 (13%) and 28 (15%) tumors classified as high, respectively. This resulted in an ORps3 for response of 0.18 and 0.16 ([95% CI, 0.06 to 0.45]; P < .001) and an ORps5 of 0.12 and 0.12 ([95% CI, 0.04 to 0.33]; P < .001) in univariate and multivariate analyses, respectively. The univariate HR for PFS of the dichotomized HOXB13-to-IL17BRps3 and IL17BRps5 ratio were 3.43 and 3.85 (both P < .001; Table 3). When added to the multivariate base model, high levels of HOXB13-to-IL17BR ratios were independently associated with a poor PFS (Table 3; both P < .001). We also evaluated a previously defined cutoff point of a HOXB13-to-IL17BR index for tamoxifen response in an adjuvant setting by Ma et al.7 This cutoff point (0.06) classified 65 (34%) and 73 (38%) of the tumors as high for HOXB13-to-IL17BRps3 and IL17BRps5 indexes, respectively. In univariate analysis, the OR for response of the HOXB13-to-IL17BR index based on this predefined cutoff point was ORps3 = 0.59 (95% CI, 0.32 to 1.09; P = .09) and ORps5 = 0.42 (95% CI, 0.23 to 0.76; P = .004), and it resulted in a significant association with PFS (ie, HRps3 = 2.15 and HRps5 = 2.39; Table 3). In multivariate analysis, the two-gene index remained only significantly associated with a shorter PFS (Table 3). The predictive value of HOXB13-to-IL17BRps3 was visualized with Kaplan-Meier curves in Figure 2. Similar results were obtained for the HOXB13-to-IL17BRps5 ratio.
Ma et al2 used microarrays to analyze ER-positive tumors from patients treated with adjuvant tamoxifen. They identified HOXB13 and IL17BR as being differentially expressed between relapsed patients and disease-free patients. However, in the adjuvant setting, one cannot discriminate between tumor aggressiveness and response to treatment.6 The data were confirmed with qRT-PCR on FFPE-derived RNA in a small patient cohort that was not similar to the original training set. It is generally accepted that RNA from FFPE samples is significantly degraded and, therefore, of lower quality compared to RNA from frozen samples. This article describes expression levels of HOXB13 and IL17BR measured with qRT-PCR on frozen tissue–derived RNA in a large number of tumors. Based on this technical approach and the large sample size, a HOXB13-to-IL17BR expression ratio can be more accurately quantified and related to outcome and clinicopathologic data. Up to now, this is the only study that allows for a relation of HOXB13-to-IL17BR expression levels with both tumor aggressiveness and response to first-line tamoxifen monotherapy. Our study supports the finding that a HOXB13-to-IL17BR ratio has prognostic value in LNN-untreated patients.7,8 In contrast to others,4,13 we found in a larger cohort of LNP-untreated patients a statistically significant association with disease outcome in univariate analysis. Interestingly, our findings demonstrated for the first time the relationship of a HOXB13-to-IL17BR ratio with response to first-line tamoxifen monotherapy. We showed in patients with recurrent disease that high HOXB13-to-IL17BR ratio levels were associated with a poor response to therapy and a short PFS, independent of traditional clinical and pathologic predictive factors. For their analysis of IL17BR expression levels, Reid et al4 used a primer set in the 5' end region of IL17BR, whereas Ma et al2 applied a primer set in the 3' end region. Because of the difference in primer design, we have evaluated both primer sets. The 3' end primer set (ps3) of IL17BR revealed 6 x higher expression levels compared with levels determined with a 5' end primer set (ps5). Despite this six-fold difference, levels of both correlated significantly. As a consequence, corresponding HOXB13-to-IL17BR ratios showed equivalent performances. The HOXB13 gene is localized at the edge of a HOXB-gene cluster at chromosome 17q21 and belongs to the canonical family of homeobox (HOX) genes.14 HOX proteins often require other homeodomain proteins to form DNA-binding complexes.15,16 Our data revealed that in 36% of breast tumors, HOXB13 expression was below detection level. The observed absence of HOXB13 expression in these tumors may be caused by a chromosomal deletion of the gene but also may be due to epigenetic silencing, such as promotor methylation. In renal cell carcinoma, complete methylation of 5' CpG islands of HOXB13 and corresponding loss of mRNA and protein expression were reported to correlate with tumor progression.17 In contrast, HOXB13 was absent in normal breast and prostate tissue, whereas elevated levels were detected in breast and prostate tumor tissue.2,18,19 In prostate cancer cell lines, it was shown that HOXB13 functions as an androgen receptor (AR) repressor to modulate hormone-activated AR signaling and, in contrast with observations in tumor tissue, to suppress cell growth.15 Likewise, we and others found that the absence of HOXB13 expression was predominantly observed (85%) in ER-positive breast tumors. This relation between ER expression and transcriptional repression of HOXB13 needs further investigation. In conclusion, this retrospective qRT-PCR study provides evidence that high HOXB13-to-IL17BR expression levels are associated with both tumor aggressiveness and tamoxifen monotherapy failure.
The authors indicated no potential conflicts of interest.
Conception and design: Maurice P.H.M. Jansen, Els M.J.J. Berns Provision of study materials or patients: Jan G.M. Klijn, John A. Foekens Collection and assembly of data: Anieta M. Sieuwerts, Kirsten Ritstier, Iris L. van Staveren Data analysis and interpretation: Maurice P.H.M. Jansen, Anieta M. Sieuwerts, Maxime P. Look, Marion E. Meijer-van Gelder Manuscript writing: Maurice P.H.M. Jansen, Maxime P. Look, Els M.J.J. Berns Final approval of manuscript: Maurice P.H.M. Jansen, Anieta M. Sieuwerts, Maxime P. Look, Jan G.M. Klijn, John A. Foekens, Els M.J.J. Berns
We thank the patients, surgeons, pathologists, and internists of the St Clara Hospital, Ikazia Hospital, St Fransiscus Gasthuis, Erasmus MC at Rotterdam, and Ruwaard van Putten Hospital at Spijkenisse for their assistance in collecting tumor tissues and patients' clinical follow-up data. We also thank Mieke Timmermans, Anita Trapman, Roberto Rodriguez Garcia, Miranda Arnold, Anneke Goedheer, Vanja de Weerd, and Henk Portengen for their technical support.
Supported by Grant No. DDHK 2001-2364 of the Dutch Cancer Society, Amsterdam, the Netherlands, and in part by the Netherlands Genomics Initiative/ Netherlands Organization for Scientific Research. Presented in part at the Fifth European Breast Cancer Conference, Nice, France, March 21-25, 2006. Both M.P.H.M.J. and A.M.S. contributed equally to this article. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Early Breast Cancer Trialists' Collaborative Group (EBCTCG): Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: An overview of the randomised trials. Lancet 365:1687-1717, 2005[CrossRef][Medline] 2. Ma XJ, Wang Z, Ryan PD, et al: A two-gene expression ratio predicts clinical outcome in breast cancer patients treated with tamoxifen. Cancer Cell 5:607-616, 2004[CrossRef][Medline] 3. Jansen MP, Foekens JA, van Staveren IL, et al: Molecular classification of tamoxifen-resistant breast carcinomas by gene expression profiling. J Clin Oncol 23:732-740, 2005 4. Reid JF, Lusa L, De Cecco L, et al: Limits of predictive models using microarray data for breast cancer clinical treatment outcome. J Natl Cancer Inst 97:927-930, 2005 5. Fan C, Oh DS, Wessels L, et al: Concordance among gene-expression-based predictors for breast cancer. N Engl J Med 355:560-569, 2006 6. Jansen MP, Foekens JA, Klijn JG, et al: Re: Limits of predictive models using microarray data for breast cancer clinical treatment outcome. J Natl Cancer Inst 97:1851-1853, 2005 7. Ma X, Hilsenbeck S, Wang W, et al: The HOXB13: IL17BR expression index is a prognostic factor in early stage breast cancer. J Clin Oncol 24:4611-4619, 2006 8. Erlander MG, Ma XJ, Hilsenbeck SG, et al: Validation of HOXB13, IL17BR and CHDH as predictors of clinical outcome of adjuvant tamoxifen monotherapy in breast cancer. Breast Cancer Res Treat 94:S33-S34, 2005 9. Hayward JL, Carbone PP, Heuson JC, et al: Assessment of response to therapy in advanced breast cancer: A project of the Programme on Clinical Oncology of the International Union Against Cancer, Geneva, Switzerland. Cancer 39:1289-1294, 1977[CrossRef][Medline] 10. Foekens JA, Peters HA, Grebenchtchikov N, et al: High tumor levels of vascular endothelial growth factor predict poor response to systemic therapy in advanced breast cancer. Cancer Res 61:5407-5414, 2001 11. Sieuwerts AM, Meijer-van Gelder ME, Timmermans M, et al: How ADAM-9 and ADAM-11 differentially from estrogen receptor predict response to tamoxifen treatment in patients with recurrent breast cancer: A retrospective study. Clin Cancer Res 11:7311-7321, 2005 12. Meijer-van Gelder ME, Look MP, Peters HA, et al: Urokinase-type plasminogen activator system in breast cancer: Association with tamoxifen therapy in recurrent disease. Cancer Res 64:4563-4568, 2004 13. Goetz MP, Suman VJ, Ingle JN, et al: A two-gene expression ratio of homeobox 13 and interleukin-17B receptor for prediction of recurrence and survival in women receiving adjuvant tamoxifen. Clin Cancer Res 12:2080-2087, 2006 14. Grier DG, Thompson A, Kwasniewska A, et al: The pathophysiology of HOX genes and their role in cancer. J Pathol 205:154-171, 2005[CrossRef][Medline] 15. Jung C, Kim RS, Zhang HJ, et al: HOXB13 induces growth suppression of prostate cancer cells as a repressor of hormone-activated androgen receptor signaling. Cancer Res 64:9185-9192, 2004 16. Jung C, Kim RS, Lee SJ, et al: HOXB13 homeodomain protein suppresses the growth of prostate cancer cells by the negative regulation of T-cell factor 4. Cancer Res 64:3046-3051, 2004 17. Okuda H, Toyota M, Ishida W, et al: Epigenetic inactivation of the candidate tumor suppressor gene HOXB13 in human renal cell carcinoma. Oncogene 25:1733-1742, 2006[CrossRef][Medline] 18. Cantile M, Pettinato G, Procino A, et al: In vivo expression of the whole HOX gene network in human breast cancer. Eur J Cancer 39:257-264, 2003[CrossRef][Medline] 19. Edwards S, Campbell C, Flohr P, et al: Expression analysis onto microarrays of randomly selected cDNA clones highlights HOXB13 as a marker of human prostate cancer. Br J Cancer 92:376-381, 2005[CrossRef][Medline] Submitted May 10, 2006; accepted November 29, 2006.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
