Expression and splicing of the unfolded protein response gene XBP-1 are significantly associated with clinical outcome of endocrine-treated breast cancer Michael P.A. Davies 1 * , Dong Liu Barraclough 2 , Ceri Stewart 3 , Kathryn A. Joyce 1 , Richard M. Eccles 1 , Roger Barraclough 3 , Philip S. Rudland 2,3 and David Ross Sibson 1 1 Division of Surgery and Oncology, School of Cancer Studies, University of Liverpool, Liverpool, United Kingdom 2 Cancer Tissue Bank Research Centre, University of Liverpool, Liverpool, United Kingdom 3 School of Biological Sciences, University of Liverpool, Liverpool, United Kingdom X-box binding protein 1 (XBP-1) is stimulated by endoplasmic reticulum stress as part of the unfolded protein response (UPR), which can promote apoptosis or cell survival. Non-conventional splicing, stimulated during the UPR, converts mRNA for ‘‘unspliced’’ XBP-1U to ‘‘spliced’’ XBP-1S mRNA. XBP-1 mRNA is oestrogen-responsive, but XBP-1S confers oestrogen independ- ence and anti-oestrogen resistance to breast cancer cell lines. We therefore evaluated XBP-1 mRNA splicing as a factor in response of breast cancer patients to endocrine treatment. XBP-1 isoforms were measured by quantitative RT-PCR in 100 primary breast cancer patients treated with adjuvant tamoxifen (including 30 ERa-negative cases). In ERa-positive cases, levels of XBP-1U mRNA correlated with ERa mRNA levels and were lower in grade 3 tumors. Higher levels of XBP-1U mRNA were significantly asso- ciated with breast cancer survival (Log-rank p 5 0.002; Cox haz- ard ratio (HR) 0.2, p 5 0.005), independent of grade, size, nodal status and progesterone receptor status. However, in the full cohort, higher ratios of XBP-1S/XBP-1U mRNA (indicating enhanced splicing) were associated with poor survival (Log-rank p 5 0.03; Cox HR 2.3, p 5 0.03) and related factors: ERa-negative status, progesterone receptor negative status, grade 3 tumors and greater proliferation. Significant associations with poor outcome were also seen for XBP-1 splicing in ERa-positive cases. Our find- ings, that XBP-1 isoforms are differently associated with outcome of endocrine therapy for patients, can be explained by higher lev- els of dominant-negative XBP-1U favouring apoptosis of tumor cells and higher levels of XBP-1S increasing tumor survival. ' 2008 Wiley-Liss, Inc. Key words: breast cancer; unfolded protein response; XBP-1; tamoxifen; outcome Mechanisms of resistance to adjuvant endocrine treatment are a significant clinical problem and are not fully known or under- stood. 1 The presence of ERa, especially in association with other ERa-responsive genes (e.g., progesterone receptor) is used to identify breast cancers more likely to respond to anti-oestrogen therapy. 1 X-box binding protein 1 (XBP-1) is found in a variety of tumors 2 including breast cancers, 3,4 where expression profiling has identified XBP-1 mRNA as being associated with oestrogen receptor a (ERa). 5 XBP-1 mRNA is oestrogen-responsive in nor- mal breast tissue 6 and breast cancer cell lines. 7 In addition, the protein transactivates ERa in both a ligand-dependent and inde- pendent manner. 8 It has recently been found 9 that over-expression of the spliced form of XBP-1 (XBP-1S) confers to breast cancer cell-lines, both oestrogen-independence and resistance to 2 differ- ent classes of anti-oestrogen, represented by tamoxifen and fulves- trant. However, little is known about the significance of XBP-1 splicing in breast cancers. XBP-1 is a CREB/ATF transcription factor. 10,11 It acts during the unfolded protein response (UPR) and can increase tumor-cell survival by inhibiting apoptosis, 12 for example, under hypoxic conditions. 13 The UPR is an adaptive response to endoplasmic reticulum stress, signalled by the accumulation of incorrectly folded proteins, for example, during hypoxia or nutrient depriva- tion. 11 ATF6 stimulates XBP-1 expression and IRE1 mediates non-conventional splicing of ‘‘unspliced’’ XBP-1 mRNA (XBP- 1U) to remove 26 nucleotides, 10,11 giving rise to ‘‘spliced’’ XBP-1 mRNA (XBP-1S). Both mRNAs are translated to protein, but a frame-shift in XBP-1S mRNA results in a larger transcription fac- tor with altered properties. 11 XBP-1S is considered to be the more active form transcriptionally, but it is likely that XBP-1U protein is also functional. 14 Normally, XBP-1U protein is rapidly de- graded, 15 but can bind to XBP-1S leading to its concomitant deg- radation, thus modulating the UPR. 16 Such dominant negative inhibition can inhibit the adaptive response to endoplasmic reticu- lum stress and ultimately lead to apoptosis. The relative levels of XBP-1 isoforms may, therefore, influence the fate of cancer cells under stress. Here, quantitative reverse transcription-PCR (qRTPCR) has been used to show that differential expression and splicing of XBP-1 mRNA in primary breast cancer is a significant marker for the outcome of adjuvant endocrine-treatment of breast cancer. Material and methods Patients This was a retrospective cohort study of 100 breast cancer cases from the Liverpool Cancer Tissue Bank Research Centre, 17 details in Table I. All were post-menopausal women (median age 67 years), who underwent mastectomy (n 5 34) or wide local exci- sion (n 5 66), some with radiotherapy (n 5 50). All received adju- vant anti-oestrogen treatment (mostly tamoxifen, 93 cases), but no primary endocrine treatment or chemotherapy. ERa, progesterone receptor (PgR) and Ki67 status were reported previously. 17,18 Clinical follow-up data were recorded by retrospective case-note review. Outcome measures were relapse-free survival (RFS) and breast cancer survival (BCS). Median follow-up was 69 months for RFS and 75 months for BCS (range 11–112). Ethical approval for the study was obtained from all relevant bodies. qRTPCR Total RNA (0.5 lg) underwent reverse transcription in dupli- cate with Superscript II Reverse Transcriptase (Invitrogen, Paisely, UK) according to manufacturer’s instructions. PCR reac- tions were performed in a Bio-Rad Icycler Real-Time PCR machine. Non-splice-specific flanking primers (upper: 5 0 -AAGC- CAAGGGGAATGAAGT-3 0 , lower: 5 0 -CCAGAATGCCCAA- CAGGATA-3 0 ) amplified the portion of XBP-1 incorporating the 26nt ‘‘exon’’ and splice-specific Taqman probes detected each iso- Grant sponsors: Clatterbridge Cancer Research Trust, Liverpool Cancer Tissue Bank Research Centre. *Correspondence to: Clatterbridge Cancer Research Trust, JK Douglas Laboratories, Clatterbridge Hospital, Wirral, Merseyside, CH63 4JY, UK. Fax: 144-151-3431820. E-mail: Michael.Davies@liverpool.ac.uk Received 17 September 2007; Accepted after revision 10 January 2008 DOI 10.1002/ijc.23479 Published online 3 April 2008 in Wiley InterScience (www.interscience. wiley.com). Abbreviations: BCS, breast cancer survival; CI, 95% confidence inter- vals; ERa, oestrogen receptor alpha; HR, hazard ratios; PgR, progesterone receptor; qRTPCR, quantitative reverse transcription-PCR; RFS, relapse- free survival; ROC, receiver operating characteristic; UPR, unfolded pro- tein response; XBP-1, X-box binding protein 1; XBP-1S, ‘‘spliced’’ XBP- 1; XBP-1U,‘‘unspliced’’ XBP-1. Int. J. Cancer: 123, 85–88 (2008) ' 2008 Wiley-Liss, Inc. Publication of the International Union Against Cancer