Abstract
The most important clinical biomarker for breast cancer management is oestrogen receptor alpha (ERα). Tumours that express ER are candidates for endocrine therapy and are biologically less aggressive, while ER-negative tumours are largely treated with conventional chemotherapy and have a poor prognosis. Despite its significance, the mechanisms regulating ER expression are poorly understood. We hypothesised that the inflammatory cytokine oncostatin M (OSM) can downregulate ER expression in breast cancer. Recombinant OSM potently suppressed ER protein and mRNA expression in vitro in a dose- and time-dependent manner in two human ER+ breast cancer cell lines, MCF7 and T47D. This was dependent on the expression of OSM receptor beta (OSMRβ) and could be blocked by inhibition of the MEKK1/2 mitogen-activated protein kinases. ER loss was also necessary for maximal OSM-induced signal transduction and migratory activity. In vivo, high expression of OSM and OSMR mRNA (determined by RT-PCR) was associated with reduced ER (P<0.01) and progesterone receptor (P<0.05) protein levels in a cohort of 70 invasive breast cancers. High OSM and OSMR mRNA expression was also associated with low expression of ESR1 (ER, P<0.0001) and ER-regulated genes in a previously published breast cancer gene expression dataset (n=321 cases). In the latter cohort, high OSMR expression was associated with shorter recurrence-free and overall survival in univariate (P<0.0001) and multivariate (P=0.022) analyses. OSM signalling may be a novel factor causing suppression of ER and disease progression in breast cancer.
Introduction
Oestrogen receptor alpha (ERα) is a central factor in breast cell biology and growth. As the primary transcription factor that mediates oestrogen signalling, ER is the linchpin of endocrine therapy and a feature that partly defines the various molecular breast cancer subtypes (Sotiriou & Pusztai 2009, Prat et al. 2010). While the mechanisms that regulate ER expression are clearly important for our understanding and clinical management of breast cancer, these remain poorly characterised.
Endocrine therapy is highly effective for the treatment of ER+ breast cancer (EBCTCG 2005). However, ∼30% of primary tumours are ER− at clinical presentation, constituting the principal mechanism of intrinsic resistance to endocrine therapy (Musgrove & Sutherland 2009). Among the tumours that are ER+ and respond initially to endocrine therapy, some develop acquired resistance through the suppression of ER expression, a mechanism that may account for up to 20% of resistant breast cancers (Musgrove & Sutherland 2009). Mechanisms to explain ER suppression include hypermethylation of the ER gene promoter (Stearns et al. 2007), destabilisation of ER mRNA (Reid et al. 2002), hypoxia (Cooper et al. 2004) and hyperactivation of mitogen-activated protein kinase (MAPK) signalling (Oh et al. 2001, Creighton et al. 2006, 2008, Bayliss et al. 2007, Lopez-Tarruella & Schiff 2007). Prolonged growth of breast cells as mammospheres or upregulation of the transcription factors snail and slug may also cause ER suppression (Dhasarathy et al. 2007, Storci et al. 2010, Guttilla et al. 2012).
Leucocytes can influence malignant cells through various mechanisms, particularly cytokine release (Balkwill et al. 2005), and ER− breast tumours are generally enriched for intra-tumoral leucocytes compared with ER+ lesions (Teschendorff et al. 2007a). This raises the possibility that tumour-associated immune responses could influence ER expression. As many of the processes shown to influence ER are highly dynamic, the ER negativity of some tumours, whether observed at diagnosis or upon disease progression, might be reversible.
We and others have shown that oncostatin M (OSM), a cytokine of the interleukin 6 (IL6) family, promotes acquisition of aggressive features such as enhanced migration and invasiveness (Zhang et al. 2003, Holzer et al. 2004, Jorcyk et al. 2006, Underhill-Day & Heath 2006, West & Watson 2010). OSM is produced by leucocytes including T cells, monocytes and neutrophils (Tanaka & Miyajima 2003, Queen et al. 2005) and engages heterodimeric receptors involving gp130 and either the OSM receptor beta (OSMRβ) or the leukaemia inhibitory factor receptor (LIFR). Signal transduction (Heinrich et al. 2003) is initiated by Janus kinases (JAKs) that engage the MAPK, signal transducer and activator of transcription 3 (STAT3) and phosphatidylinositol-3-kinase (PI3K (PIK3CA)) pathways, each of which has documented roles in breast tumour pathogenesis (Clevenger 2004, Dillon et al. 2007, Whyte et al. 2009). Based on the emerging recognition that immune activity can affect breast cancer biology (DeNardo & Coussens 2007), and a prior observation that OSM may influence ER in MCF7 cells (Grant et al. 2002), our primary aim in this study was to assess OSM signalling as a possible novel regulator of ER expression in breast cancer.
Materials and methods
Cell culture and cytokine stimulation
Human breast carcinoma cell lines MCF7, T47D and ZR75-1 (ATCC, Manassas, VA, USA) were cultured in DMEM with 5% FBS under standard conditions. Human OSM, IL6 and TNF-α (Peprotech, Rocky Hill, NJ, USA) were stored as 100 μg/ml stocks in culture media and, unless otherwise specified, used at 100 ng/ml.
Chemical inhibitors and RNA interference
Inhibitors to MEKK1/2 (U0126 (Favata et al. 1998) and PD98059 (Dudley et al. 1995); Cell Signaling, Danvers, MA, USA), JAKs (JAK inhibitor I (Thompson et al. 2002)), Ras (FTI277 (Vogt et al. 1996)), c-jun N-terminal kinase (JNK) (JNK inhibitor VIII (Szczepankiewicz et al. 2006)), p38 MAPK (SB203580 (Cuenda et al. 1995)), PI3K (LY294002 (Vlahos et al. 1994)), EGFR (AG1478 (Osherov & Levitzki 1994)), mTOR (rapamycin) or NF-κB (oridonin (Ikezoe et al. 2005) (Calbiochem, San Diego, CA, USA) were added to cultures 30 min before cytokine stimulation at doses of 10 μM, with the exception of PD98059 (50 μM), rapamycin (10 nM) and oridonin (10 μg/ml). All inhibitors used in this study are established reagents for specific and effective disruption of their respective targets. Although the JAK inhibitor may suppress MAPK activity at the concentration used in this study, this was addressed by including it in all inhibition experiments as a positive control reagent for suppression of all aspects of OSM signalling. Gene knockdown was performed using ON-TARGET plus SMARTpool siRNA at a final concentration of 100 nM transfected with Dharmafect-4 (Dharmacon, Lafayette, CO, USA).
Western blots
Cells were prepared for immunoblotting as described previously (Al-Haddad et al. 1999). Protein concentrations were estimated using an ND-1000 spectrophotometer (NanoDrop, Wilmington, DE, USA). Primary antibodies were ERα (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA, USA), GAPDH (1:3000; Stem Cell Technologies, Vancouver, BC, Canada), β-actin (1:3000; Abcam, Cambridge, MA, USA), phospho-STAT3 (Tyr705; 1:1000), phospho-AKT (Ser473; 1:500), phospho-ERK1/2 (Thr202/Tyr204; 1:500) and progesterone receptor (PR; 1:500; Cell Signaling). Secondary antibodies were HRP-conjugated bovine anti-rabbit and goat anti-mouse IgG (1:3000; Santa Cruz Biotechnology). STAT3 phosphorylation was used throughout this study as an indicator of OSM functionality (Kan et al. 2011). Band densitometry was performed using ImageJ.
Real-time quantitative PCR (RT-PCR)
Total RNA was extracted from cell lines and tissues using the Qiagen RNEasy mini kit (Qiagen) and quantified using an ND-1000 spectrophotometer. RNA was reverse transcribed using the qScript cDNA synthesis kit (Quanta Biosciences, Gaithersburg, MD, USA). RT-PCR was performed using Perfecta SYBR Green supermix (Quanta Biosciences) and an iCycler thermal cycler with a MyIQ real-time PCR detection system (Bio-Rad). Reactions were performed in triplicate for each sample. Data for all target genes were normalised to RPL27 expression. Intron-spanning primers were purchased from Integrated DNA Technologies (IDT, Coralville, IA, USA) and designed to have annealing temperatures of 60 °C. Primer sequences are as follows (listed 5′–3′): RPL27 – forward, CAATCACCTAATGCCCACAAG; reverse, TTCTTGCCTGTCTTGTATCTCTC; ESR1 – forward, CGACTATATGTGTCCAGCCAC; reverse, CCTCTTCGGTCTTTTCGTATCC; PGR – forward, TCGCCTTAGAAAGTGCTGTC; reverse, GCTTGGCTTTCATTTGGAACG; OSM – forward, CTCGAAAGAGTACCGCGTG; reverse, TCAGTTTAGGAACATCCAGGC; and OSMR – forward, TCCCAATACCACAAGCACAG; reverse, GCAAGTTCCTGAGAGTATCCTG.
ER functional assays
To assay responses to 17β-estradiol (E2, 10 nM), cells were cultured in E2-free media (phenol red-free DMEM with 5% charcoal-dextran-stripped FBS) for 72 h before further treatment. To assess the effect of ER on cell migration, MCF7 cells were transfected with pcDNA3.1 or pcDNA3.1-ERα using Fugene 6 (Roche) at a Fugene:DNA ratio of 9 μl:1 μg. Two days later, cells were split into control or OSM treatment groups and on the following day seeded in triplicate onto 8 μm pore polycarbonate filters in 24-well plates at a density of 50 000 cells/well using FBS as the chemoattractant. Cells were fixed 24 h later in 3.7% formaldehyde and stained with crystal violet. Cells on the upper membrane surface were removed with a cotton swab and migrated cells in four random fields counted under high power (200×) and averaged to produce one of the three replicate values.
Clinical cohorts
Two independent cohorts of human breast cancer were used. The first included 70 invasive breast carcinomas obtained from the Manitoba Breast Tumour Bank (MBTB), which operates with approval from the Research Ethics Board of the Faculty of Medicine, University of Manitoba. Cases were selected to represent several molecular subtypes (luminal-A, luminal-B, Her2 and basal-like and triple-negative non-basal (TNNB)). ER and PR status were previously determined by ligand-binding assay (LBA) and Her2, EGFR, CK5/6 and Ki67 status were determined by immunohistochemistry in previous studies (Skliris et al. 2008, Blanchard et al. 2009). Luminal-A and -B tumours were defined as ER+ and/or PR+ (LBA scores of ≥10 and >15, respectively) and Her2– (IHC score <3+); luminal-B tumours were additionally Ki67+ (>10%+). Her2 tumours had an IHC clinical score of 3+. Basal-like tumours were triple negative for ER, PR and Her2 and were positive for CK5/6 and/or EGFR (IHC scores derived from the product of intensity and % positivity, as described previously (Wang et al. 2008)). TNNB tumours were triple negative but lacked CK5/6 and EGFR expression. Frozen tissue sections were used for RNA extraction and RT-PCR. Additional frozen tumour samples used to generate supplementary data were obtained from the BC Cancer Agency's Tumor Tissue Repository (REB certificate #H06-60001).
The second cohort was derived from a publically accessible microarray gene expression dataset (Prat et al. 2010), acquired from the University of North Carolina Microarray Database (UNC). Data from probes matching the same gene were collapsed by averaging (information regarding assay platforms and individual probes can be obtained from the Gene Expression Omnibus web site using accession number
Statistical analysis
All experiments were repeated at least thrice unless otherwise specified. Specific tests are specified in figure and table legends; all were two sided with significance established at <0.05. Univariate tests were performed using Prism 5.0 (GraphPad, La Jolla, CA, USA) and multivariate tests with Statistics 14 (SPSS, Chicago, IL, USA).
Results
OSM signalling suppresses ER expression
We began by studying the effect of OSM on three ER+ human breast cancer cell lines. Stimulation of both MCF7 and T47D cells with escalating doses of OSM revealed maximal suppression of ER at a dose of 100 ng/ml after 24 h of treatment. This corresponded to a reduction in ER protein by up to 95% in MCF7 cells and 85% in T47D cells (Fig. 1A). In contrast, OSM had no apparent effect on ER expression in a third ER+ cell line, ZR75-1 (data not shown). ESR1 (ER) mRNA levels in MCF7 cells were also significantly reduced after 24 h of OSM stimulation, as were those of PGR (PR), a key target gene of ER transcriptional activity (Fig. 1B). Unlike OSM, stimulation with IL6, the prototypic cytokine of the OSM family, had comparatively little impact on ER expression (Fig. 1C). Time-course assays revealed that although OSM caused rapid (<1 h) activation of signalling effectors such as STAT3 and ERK1/2, ER protein levels did not noticeably diminish until after ∼6 h of OSM treatment. With ongoing stimulation, ER levels remained stably suppressed for at least 96 h (Fig. 1D).
Suppression of ER by OSM depends on expression of OSMR
As OSM can engage both OSMR-gp130 and LIFR-gp130 receptor heterodimers, we assessed the degree of receptor specificity for the ER-suppressive activity of OSM by knocking down OSMR expression with siRNA. In MCF7 cells transfected with a control GFP-targeted siRNA, 24 h of OSM treatment caused an unexpected four- to five-fold increase in OSMR mRNA, but this was completely abrogated by transfection with OSMR siRNA. OSMR knockdown prevented suppression of both ESR1 and PGR by OSM (Fig. 2A). This was also evident at the protein level (Fig. 2C). As noted above, OSM had a robust impact on ER levels in MCF7 cells, a reduced effect on T47D cells and no observable impact on ZR75-1 cells. This is consistent with the level of OSMR expression in these cell lines; relative to MCF7 cells, T47D cells have comparable but lower levels of OSMR expression, while expression in ZR75-1 cells is nearly 100-fold lower (Fig. 2B). These data suggest that suppression of ER by OSM occurs principally via the OSMR-gp130 heterodimer.
ER suppression by OSM is reversible and dependent on MAPK signalling
To determine the persistence of ER suppression by OSM, we stimulated MCF7 and T47D cells for 48 h, withdrew OSM and cultured the cells for a further 24 h in cytokine-free media. Withdrawal of OSM caused full restoration of ER expression, indicating that the effect of OSM is transient in the absence of ongoing stimulation (Fig. 3A). To determine the signal transduction requirements for ER suppression, we individually attenuated the STAT3, PI3K and MAPK pathways before OSM treatment. As expected, JAK inhibition blocked downstream pathways and ER loss (Fig. 3B, left panel). Transfection of MCF7 cells with STAT3-specific siRNA failed to attenuate ER suppression (Fig. 3B, right panel), as did blockade of PI3K activity using LY294002 (Fig. 3B, left panel). Treatment with the MEKK1/2 inhibitor U0126, however, partly restored ER expression (Fig. 3B, left panel), along with an alternative MEKK inhibitor, PD98059, and the farnesyltransferase inhibitor FTI277, a disruptor of Ras processing and downstream MAPK activity (Supplementary Figure S1, see section on supplementary data given at the end of this article). Although OSMR physically associates and cooperates with EGFR, inhibition of EGFR with AG1478 did not affect ER nor did inhibition of the growth factor signalling mediators NF-κB and mTOR using the NF-κB DNA-binding inhibitors oridonin and rapamycin respectively (Supplementary Figure S2, see section on supplementary data given at the end of this article). Similarly, although OSM can activate other MAPKs including JNK and p38 MAPK, inhibition of these pathways had no impact on ER (Supplementary Figure S2), suggesting that OSM-induced ER suppression may be specifically due to the ERK1/2 pathway. Intriguingly, blockade of ERK1/2 in MCF7 cells also prevented the morphological changes characteristic of OSM signalling, implying that ER suppression and the gain of motility-associated morphology may be mechanistically linked (Fig. 3C).
Suppression of ER is functionally important during OSM signalling
To determine whether suppression of ER by OSM was functionally relevant, we first assessed the ability of OSM-stimulated cells to respond to the ER ligand, E2. In MCF7 and T47D cells that had been conditioned by growth in hormone-free conditions for 3 days and subsequently stimulated with E2, OSM-treated cells failed to respond to E2 by upregulating expression of PR (Fig. 4A and Supplementary Figure S3A, see section on supplementary data given at the end of this article). Cell proliferation in response to E2 was similarly inhibited by OSM treatment, though in our experimental conditions this was statistically significant only for T47D cells (Supplementary Figure S3B). To assess the role of ER suppression in OSM-induced migration, we constitutively overexpressed ER in MCF7 cells and subjected them to transwell migration assays. Control-transfected cells displayed a sixfold increase in migration 48 h following OSM treatment. In contrast, migration was only modestly enhanced by OSM in ER-transfected cells (Fig. 4B), which displayed significantly reduced activation of STAT3 and ERK1/2, but not PI3K (Fig. 4C). Collectively, these data indicate that OSM can reduce ER expression below a functionally critical threshold and, furthermore, that suppression of ER may be required for full engagement of OSM-induced signal transduction and cell migration.
The OSM pathway correlates with suppression of ER in vivo
To investigate the association of OSM signalling with ER in human tumours, we examined OSM and OSMR expression by RT-PCR in a cohort of 70 invasive breast carcinomas (MBTB cohort; see Materials and methods section). Data were analysed using the median expression values as cutpoints to produce two patient groups: those high in both OSM and OSMR and those with low levels of one or both. This grouping was chosen on the assumption that both OSM and OSMR must be expressed for full activation of OSM signalling. Overall, 12% (4/32) of ER+ tumours were associated with high OSM/OSMR status compared with 45% (17/38) of ER− tumours (Fisher's exact test, P=0.0041), and ER and PR protein levels were reduced by seven- to eight-fold in the OSM/OSMR-high group (P<0.05, Fig. 5A). To determine whether this association was related to tumour differentiation, we compared the frequency of high OSM/OSMR between each of five major molecular subtypes of breast cancer (luminal-A, luminal-B, Her2, basal-like and TNNB subgroups). High OSM/OSMR status was rare in the luminal subtypes (<5% tumours) but was seen in 20–60% of tumours within the other subtypes. When compared with luminal-A tumours, OSM/OSMR expression was significantly more frequent within the Her2 and basal-like classes (P=0.0135 and 0.0029, respectively, Fisher's exact test, Fig. 5B). As high levels of growth factor signalling may independently contribute to suppression of ER (Lopez-Tarruella & Schiff 2007), a larger subset of Her2 subtype tumours was included within the cohort to assess the association between OSM/OSMR and ER independently of Her2 status. Within this small Her2+ subgroup, there was still a significant inverse correlation between OSM/OSMR and ER levels (Spearman r=−0.626, P=0.0014).
To validate these findings in a larger cohort, we examined data from previously published gene expression microarrays of over 300 invasive breast cancers (‘Prat cohort’ (Prat et al. 2010)). Due to the larger sample size of this dataset, we were able to consider three groups of tumours with respect to OSM/OSMR expression (using medians as cutpoints): those high in both OSM and OSMR (high/high, n=77), those high in one or the other (high/low, n=161) and those low in both (low/low, n=83). ESR1 expression was substantially reduced in the high/low and high/high groups relative to the low/low group, with a clear trend towards lower ESR1 as representation of the OSM pathway increased (Fig. 5C). This pattern was also largely replicated with respect to four ER-regulated genes: PGR (PR), trefoil factor 1 or pS2 (TFF1), GATA-binding protein 3 (GATA3) and cyclin D1 (CCND1). This implies that the ER pathway as a whole, rather than simply ER alone, is disabled in tumours with robust OSM activity. Consistent with the MBTB cohort, high OSM/OSMR expression was, relative to the luminal-A subtype, strongly associated with the Her2, basal-like and claudin-low subtypes, each of which is notable for low hormone receptor expression (Fig. 5D; P=0.0003, 0.0007 and 0.0009, respectively, χ2 test). OSM/OSMR-high status was correlated with lower ESR1 expression even when analysis was restricted to clinically ER− cases (data not shown). As with the MBTB cohort, clinically Her2+ tumours with high OSM/OSMR expression had considerably reduced ESR1 expression relative to low/low cases (P=0.0184, data not shown).
High OSMR expression is associated with poor prognosis
When the prognosis of the OSM/OSMR groups described above was assessed in the Prat cohort, we observed increased risk of recurrence in the high/low and high/high groups (P=0.0323; data not shown). However, when OSM and OSMR were assessed individually, only OSMR expression was associated with poor prognosis. To further explore this relationship, we first filtered the cohort by excluding cases with high (upper quartile) expression of the macrophage marker CD68 (to increase the probability that assessed OSMR expression was derived from malignant epithelium, as OSMR can be highly expressed in myeloid leucocytes (Dillon et al. 2004)). Cases high in OSMR (upper quartile) within this filtered cohort (total n=241) had a much greater risk of recurrence (HR=5.06, 95% CI 2.49–10.29; P<0.0001; Fig. 6A) and overall mortality (HR=5.64, 95% CI 2.63–12.11; P<0.0001). OSMR-high status was strongly associated with clinical ER− status (P<0.0001) but not clinical PR or Her2 status, nor lymph node metastasis, patient age, tumour grade or tumour size (Table 1). In multivariate Cox regression modelling of disease-free survival (DFS) involving the above parameters, OSMR was significantly associated with survival (HR=2.75, 95% CI 1.15–6.55; P=0.022) along with lymph node and PR status (Table 1) and remained significant when molecular subtypes were included in the model (P=0.025). Indeed, OSMR was strongly prognostic even in the poor outcome basal-like (P=0.0087) and Her2 (P=0.0001) intrinsic subtypes (Supplementary Figure S4, see section on supplementary data given at the end of this article). OSMR-high status was also associated with loss of co-expression of ESR1 and ER-regulated genes (P<0.0001; Fig. 6B). Among OSMR-high cases, those that retained high expression of both ESR1 and PGR had a highly favourable prognosis relative to ESR1/PGR-suppressed cases (n=41, P=0.0014; Fig. 6C). This supports the concept that suppression of ER is required for OSM to fully activate an aggressive phenotype in breast cancer cells.
Prat cohort associations between OSMR and clinical parameters and their relationship with DFS in Cox regression modellinga
Association with OSMRb | Cox association with DFS (n=132)c | |||||
---|---|---|---|---|---|---|
Parameter | OSMR low | OSMR high | P value | Comparison | HR (95% CI) | P value |
Patient age (year) | ||||||
<50 | 53 | 23 | 0.0762 | <50 vs ≥50 | 0.48 (0.22–1.05) | 0.067 |
≥50 | 90 | 20 | ||||
Tumour size | ||||||
T1 | 30 | 9 | 0.305 | T3/4 vs T2 vs T1 | 1.67 (0.87–3.20) | 0.122 |
T2 | 78 | 18 | ||||
T3–T4 | 35 | 15 | ||||
Tumour grade | ||||||
1–2 | 64 | 16 | 0.2848 | 3 vs 1–2 | 1.71 (0.72–4.08) | 0.226 |
3 | 66 | 25 | ||||
Nodal status | ||||||
Negative | 70 | 14 | 0.0805 | Pos. vs neg. | 2.85 (1.31–6.20) | 0.008 |
Positive | 74 | 29 | ||||
ER status | ||||||
Negative | 39 | 26 | <0.0001 | Pos. vs neg. | 1.00 (0.33–3.02) | 1.00 |
Positive | 101 | 16 | ||||
PR status | ||||||
Negative | 68 | 22 | 0.1258 | Pos. vs neg. | 0.28 (0.10–0.84) | 0.023 |
Positive | 71 | 12 | ||||
Her2 status | ||||||
Negative | 120 | 34 | 0.5197 | Pos. vs neg. | 1.84 (0.78–4.31) | 0.161 |
Positive | 27 | 10 | ||||
OSMRd | 182 | 58 | NA | High vs low | 2.75 (1.15–6.55) | 0.022 |
Data reflect the cohort after removal of cases with CD68 expression within the upper quartile.
Comparisons calculated by two-sided Fisher's exact or χ2 tests as appropriate.
Calculated by Cox proportional hazards regression using the enter method.
OSMR-high cases are those with OSMR expression in the upper quartile.
As the above in vivo observations could be attributable to enrichment of OSM/OSMR expression in the principally ER− basal-like, Her2 and claudin-low subtypes, we examined luminal subtype cases within the Prat cohort (>90% ER+) and observed relatively high levels of co-expression of OSM and OSMR to be associated with reductions in both ESR1 (P=0.0022) and GATA3 (P=0.0009) expression (Supplementary Figure S5, see section on supplementary data given at the end of this article), two key contributors to the luminal subtype definition (Sorlie et al. 2001). High OSM/OSMR status was also associated with poor prognosis (P<0.0001; Supplementary Figure S5), and these observations were not due to enrichment of OSM/OSMR expression in the luminal-B subtype.
OSM expression in breast tumours is associated with the innate immune compartment
Although OSM is considered as a product of leucocytes, direct evidence for this in breast cancer is lacking. To investigate this question, we examined OSM expression by RT-PCR in tumour tissues and cell lines. While OSM was readily detectable in breast tumour tissues, we did not detect expression in three breast cell lines (ER+ MCF7 and T47D cells and ER− MDA-MB-231 cells (Supplementary Figure S6A, see section on supplementary data given at the end of this article)). In the Prat cohort, we observed a significant association (P<0.0001) between OSM and expression of the pan-leucocyte marker PTPRC (CD45, Supplementary Figure S6B). Hierarchical clustering of cases with high expression of the immune cell markers CD3Z and/or CD68 (upper quartile, n=122) using genes definitive for distinct leucocyte subtypes yielded expected myeloid/antigen presenting cell (APC), T cell, B cell and cytotoxic clusters. Of these, OSM grouped clearly with the APC subset (Supplementary Figure S6C). Analysis of individual genes showed strong associations between OSM and APC markers such as toll-like receptors, CD14 and CD163 but weak, non-existent or inverse relationships with lymphocyte markers (Supplementary Figure S6D). Thus, although this remains to be proven, innate leucocytes are a probable source of OSM in breast tumours.
Discussion
We have demonstrated that OSM drives the suppression of ER in vitro through a MAPK-dependent mechanism in breast cell lines and that this pathway is part of the pro-migratory phenotype stimulated by OSM in breast cancer. In human breast tumours, the OSM pathway is associated with reduced ER activity and poor prognosis. OSM signalling may therefore be a novel mechanism underlying ER suppression in breast cancer.
The phenotypic heterogeneity of breast cancer, of which ER expression is a key component, is currently explained by two main hypotheses. The first is a lineage-based model in which specific cells of origin (for example, luminal progenitor cells) are mutated and progress to malignancy along defined paths. The second invokes a stochastic model, whereby tumour cells are phenotypically plastic and adjust their behaviour in response to both mutagenic events and shifting environmental conditions. As an explanation for the evolution of ER− breast cancer, our data fit well with the latter hypothesis. Despite their luminal phenotype, MCF7 and T47D cells rapidly downregulate ER upon stimulation with OSM and adopt features typically associated with ER− disease, such as enhanced migration. Suppression of ER and oestrogen responsiveness may explain early observations that OSM reduces proliferation of breast tumour cells in vitro (Grant & Begley 1999). Upon withdrawal of OSM, hormone receptor suppression is reversible. These findings suggest that breast tumour cells in vivo may respond to external factors such as OSM by adopting a phenotype that, through activation of pathways such as MAPK, PI3K and STAT3, could afford them key advantages including improved survival, invasion and dissemination, independent of hormones. Upon cessation of stimulus, these cells could restore expression of ER. Such a model could explain the association between leucocyte infiltration and ER negativity, a phenomenon that is not currently understood (Teschendorff et al. 2007a). Furthermore, such a model would imply that some tumours characterised as low ER+ or ER− are in fact tumours in which dynamic cell-extrinsic factors serve to suppress ER expression. Indeed, we observed here that ER+ luminal-type breast tumours enriched for OSM and OSMR had lower ER expression and a prognosis similar to that of clinically ER− tumours. It should be noted that our data do not demonstrate that ER+ intrinsic subtypes can evolve directly into ER− subtypes due to OSM. Rather, because OSM signalling appears to be generally associated with depressed ER activity and poor prognosis in all subtypes, our data should only be interpreted as a demonstration that OSM signalling may be one among several possible mechanisms underlying ER suppression in breast cancer, regardless of intrinsic subtype.
The expression of OSMR varies among breast cancer cell lines and may explain much of the variation in their responsiveness to OSM. Compared with MCF7 cells, T47D cells had roughly half the level of OSMR mRNA (with a corresponding 30% reduction in the effect of OSM on ER expression), while ZR75 cells had nearly 100-fold less OSMR expression and no observable loss of ER in response to OSM (Fig. 2). Consistent with this notion is the observation that OSMR expression in breast tumours was more closely associated with clinical outcome than OSM. The reasons for variability in OSMR expression in breast tumours are not clear at this time but may be related to promoter methylation (Deng et al. 2009, Kim et al. 2009). Alternatively, the ability of OSM to induce OSMR expression (Fig. 2) suggests that varying OSMR levels in vivo may reflect local OSM concentrations in the tumour microenvironment. It has also been recently shown that c-myc status serves as a determinant of the net cellular response to OSM in breast cell lines (Kan et al. 2011).
ER suppression following OSM stimulation was, in our experiments, dependent on Ras-mediated MAPK signalling, with no apparent requirement for STAT3 or PI3K activity. The involvement of MAPK signalling is consistent with the data from other studies. For example, MCF7 cells engineered to overexpress EGFR or constitutively active Her2, Raf or MEK exhibited oestrogen-independent growth, suppression of ER (Oh et al. 2001) and expression of a consistent set of MAPK-regulated genes that could accurately predict ER expression in human tumours (Creighton et al. 2006). Furthermore, ER suppression in this model was reversible and could be counteracted using MAPK inhibitors (Bayliss et al. 2007). Our work expands on these studies by identifying a physiologically relevant cytokine that potently activates MAPK-dependent ER suppression.
While various leucocyte subtypes are known to produce OSM, evidence that OSM is produced by tumour cells is inconclusive. A single study of its expression in breast cancer tissue was restricted to immunohistochemistry rather than mRNA expression (Garcia-Tunon et al. 2008). Our preliminary investigation of this issue indicated that OSM is absent in at least three commonly used breast cancer cell lines and that its expression in vivo correlates strongly with markers of innate leucocytes. Intriguingly, a recent study demonstrated that macrophage-conditioned media could suppress ER expression in MCF7 cells in a MAPK-dependent manner (Stossi et al. 2011). However, the specific factors mediating this observation were not identified. Therefore, OSM produced by tumour infiltrating leucocytes may constitute a novel cell-extrinsic mechanism of ER suppression. Further studies involving in situ hybridisation for OSM mRNA localisation in breast tissues and/or direct analysis of different cell types in fresh breast tumour specimens will be required to conclusively resolve this issue.
Relative to ER+ lesions, ER− breast tumours are enriched in leucocytes and cytokines (Chavey et al. 2007, Teschendorff et al. 2007a). Nevertheless, little is known regarding the direct effects of cytokines on ER expression. IL6 and tumour necrosis factor alpha (TNF-α) are reported to suppress ER expression (Bhat-Nakshatri et al. 2004, D'Anello et al. 2010), but other studies have presented conflicting results (Speirs et al. 2000, Rubio et al. 2006). In our hands, IL6 had little effect on ER expression (Fig. 1C). Clinically, IL6 has not emerged as a predictor of response to endocrine therapy (Muss et al. 2007), nor has it shown great utility as a therapeutic target (Garber 2009). Preliminary data from our lab suggest that TNF-α does indeed suppress ER and that it can do so synergistically with OSM (NR West and PH Watson, unpublished observations). When we examined the expression of inflammatory cytokines and their receptors (OSM, IL6, LIF, IL1 and TNF-α) in the Prat dataset for associations with ER pathway expression and prognosis (as in Fig. 6), only OSMR and the TNF-α receptor (TNFRSF1A) associated with both parameters, though the OSMR relationships were considerably stronger (Supplementary Table S1, see section on supplementary data given at the end of this article). Thus, among inflammatory cytokine pathways, OSMR signalling may have a uniquely potent effect on breast tumour biology, for as-yet unknown reasons.
The correlation of OSMR with poor prognosis was not mirrored by its ligand, OSM, which may reflect a role for OSMR as the key limiting factor in this system. Alternatively, as a leucocyte product, the putative negative influence of OSM may be difficult to separate from the known beneficial impact of host immunity, particularly within ER− lesions (Teschendorff et al. 2007b, Rody et al. 2009). Notably, OSMR serves as a receptor for another cytokine, IL31, which is produced by activated Th2 cells (Dillon et al. 2004). Thus, IL31 could constitute an alternative OSMR ligand in tumours. The concept of innate leucocytes producing OSM and inducing aggressive changes in neighbouring tumour cells is consistent with the conventional mindset that these leucocytes are largely responsible for the various deleterious effects of anti-tumour immunity (Balkwill et al. 2005).
Future investigation of OSM as an ER modulator should involve in vivo pre-clinical models to further address the effect of OSM on tumorigenesis, ER expression and response to endocrine therapy. If such studies verify that OSM signalling promotes breast tumorigenesis, it is possible that this pathway could be targeted therapeutically. Translation to the clinic will require development of strategies for manipulation of either OSM production from macrophages or OSM reception and signalling through OSMR in tumour cells. Some existing chemotherapies are under investigation for their suppressive effects on myeloid cells (Naiditch et al. 2011) and new chemical and liposomal drug delivery systems that target macrophages are in development (Kelly et al. 2011, Needham et al. 2011). Development and clinical testing of monoclonal antibody-based therapies targeting IL6 signalling at both the ligand and the receptor level are already well advanced (Jones et al. 2011), suggesting that OSM/OSMR signalling may be a feasible therapeutic target. Several features of OSM and OSMR make them attractive potential targets. First, by engaging multiple signalling cascades that potently influence cell survival and migration, such as the STAT, MAPK and PI3K pathways (Heinrich et al. 2003), blockade of OSM signalling would impinge on each of these mechanisms. Secondly, because both gp130 and OSMR signal cooperatively with EGFR family receptors, blockade of OSMR could potentially attenuate EGFR/Her2 signalling (Grant et al. 2002). Thirdly, both antibody neutralisation of OSM and pharmacological inhibition of OSMR ligand binding would be feasible therapeutic mechanisms. Finally, because the effects of knocking out Osm and Osmr expression in mice appear mild (Fasnacht & Muller 2008), the side effects of targeting OSM signalling may be minimal. Further exploration of the OSM pathway as a potentially clinically relevant modulator of breast cancer is warranted.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-11-0326.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This work was supported by the Canadian Breast Cancer Foundation, BC/Yukon Region and BC Cancer Foundation. N R West is supported by a US DOD breast cancer research programme pre-doctoral award (W81XWH-08-1-0781).
Author contribution statement
N R West and P H Watson designed the study and drafted the manuscript. N R West performed the experiments. N R West, L C Murphy and P H Watson contributed to manuscript revision and approved the final version.
Acknowledgements
We thank Dr Jill Murray and Michelle Parisien for helpful discussion and technical assistance. This study was supported by the Manitoba Breast Tumour Bank and the BC Cancer Agency's Tumor Tissue Repository. We are also grateful to the authors of the Prat cohort for making their data publically available.
References
Al-Haddad S, Zhang Z, Leygue E, Snell L, Huang A, Niu Y, Hiller-Hitchcock T, Hole K, Murphy LC & Watson PH 1999 Psoriasin (S100A7) expression and invasive breast cancer. American Journal of Pathology 155 2057–2066. doi:10.1016/S0002-9440(10)65524-1.
Balkwill F, Charles KA & Mantovani A 2005 Smoldering and polarized inflammation in the initiation and promotion of malignant disease. Cancer Cell 7 211–217. doi:10.1016/j.ccr.2005.02.013.
Bayliss J, Hilger A, Vishnu P, Diehl K & El-Ashry D 2007 Reversal of the estrogen receptor negative phenotype in breast cancer and restoration of antiestrogen response. Clinical Cancer Research 13 7029–7036. doi:10.1158/1078-0432.CCR-07-0587.
Bhat-Nakshatri P, Campbell RA, Patel NM, Newton TR, King AJ, Marshall MS, Ali S & Nakshatri H 2004 Tumour necrosis factor and PI3-kinase control oestrogen receptor α protein level and its transrepression function. British Journal of Cancer 90 853–859. doi:10.1038/sj.bjc.6601541.
Blanchard AA, Skliris GP, Watson PH, Murphy LC, Penner C, Tomes L, Young TL, Leygue E & Myal Y 2009 Claudins 1, 3, and 4 protein expression in ER negative breast cancer correlates with markers of the basal phenotype. Virchows Archiv 454 647–656. doi:10.1007/s00428-009-0770-6.
Chavey C, Bibeau F, Gourgou-Bourgade S, Burlinchon S, Boissiere F, Laune D, Roques S & Lazennec G 2007 Oestrogen receptor negative breast cancers exhibit high cytokine content. Breast Cancer Research 9 R15 doi:10.1186/bcr1648.
Clevenger CV 2004 Roles and regulation of stat family transcription factors in human breast cancer. American Journal of Pathology 165 1449–1460. doi:10.1016/S0002-9440(10)63403-7.
Cooper C, Liu GY, Niu YL, Santos S, Murphy LC & Watson PH 2004 Intermittent hypoxia induces proteasome-dependent down-regulation of estrogen receptor α in human breast carcinoma. Clinical Cancer Research 10 8720–8727. doi:10.1158/1078-0432.CCR-04-1235.
Creighton CJ, Hilger AM, Murthy S, Rae JM, Chinnaiyan AM & El-Ashry D 2006 Activation of mitogen-activated protein kinase in estrogen receptor α-positive breast cancer cells in vitro induces an in vivo molecular phenotype of estrogen receptor α-negative human breast tumors. Cancer Research 66 3903–3911. doi:10.1158/0008-5472.CAN-05-4363.
Creighton CJ, Massarweh S, Huang S, Tsimelzon A, Hilsenbeck SG, Osborne CK, Shou J, Malorni L & Schiff R 2008 Development of resistance to targeted therapies transforms the clinically associated molecular profile subtype of breast tumor xenografts. Cancer Research 68 7493–7501. doi:10.1158/0008-5472.CAN-08-1404.
Cuenda A, Rouse J, Doza YN, Meier R, Cohen P, Gallagher TF, Young PR & Lee JC 1995 SB 203580 is a specific inhibitor of a MAP kinase homologue which is stimulated by cellular stresses and interleukin-1. FEBS Letters 364 229–233. doi:10.1016/0014-5793(95)00357-F.
D'Anello L, Sansone P, Storci G, Mitrugno V, D'Uva G, Chieco P & Bonafe M 2010 Epigenetic control of the basal-like gene expression profile via interleukin-6 in breast cancer cells. Molecular Cancer 9 300 doi:10.1186/1476-4598-9-300.
DeNardo DG & Coussens LM 2007 Inflammation and breast cancer. Balancing immune response: crosstalk between adaptive and innate immune cells during breast cancer progression. Breast Cancer Research 9 212 doi:10.1186/bcr1746.
Deng G, Kakar S, Okudiara K, Choi E, Sleisenger MH & Kim YS 2009 Unique methylation pattern of oncostatin M receptor gene in cancers of colorectum and other digestive organs. Clinical Cancer Research 15 1519–1526. doi:10.1158/1078-0432.CCR-08-1778.
Dhasarathy A, Kajita M & Wade PA 2007 The transcription factor snail mediates epithelial to mesenchymal transitions by repression of estrogen receptor-α. Molecular Endocrinology 21 2907–2918. doi:10.1210/me.2007-0293.
Dillon SR, Sprecher C, Hammond A, Bilsborough J, Rosenfeld-Franklin M, Presnell SR, Haugen HS, Maurer M, Harder B & Johnston J et al. 2004 Interleukin 31, a cytokine produced by activated T cells, induces dermatitis in mice. Nature Immunology 5 752–760. doi:10.1038/ni1084.
Dillon RL, White DE & Muller WJ 2007 The phosphatidyl inositol 3-kinase signaling network: implications for human breast cancer. Oncogene 26 1338–1345. doi:10.1038/sj.onc.1210202.
Dudley DT, Pang L, Decker SJ, Bridges AJ & Saltiel AR 1995 A synthetic inhibitor of the mitogen-activated protein kinase cascade. PNAS 92 7686–7689. doi:10.1073/pnas.92.17.7686.
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 2005 1687–1717. doi:10.1016/S0140-6736(05)66544-0.
Fasnacht N & Muller W 2008 Conditional gp130 deficient mouse mutants. Seminars in Cell & Developmental Biology 19 379–384. doi:10.1016/j.semcdb.2008.07.001.
Favata MF, Horiuchi KY, Manos EJ, Daulerio AJ, Stradley DA, Feeser WS, Van Dyk DE, Pitts WJ, Earl RA & Hobbs F et al. 1998 Identification of a novel inhibitor of mitogen-activated protein kinase kinase. Journal of Biological Chemistry 273 18623–18632. doi:10.1074/jbc.273.29.18623.
Garber K 2009 First results for agents targeting cancer-related inflammation. Journal of the National Cancer Institute 101 1110–1112. doi:10.1093/jnci/djp266.
Garcia-Tunon I, Ricote M, Ruiz A, Fraile B, Paniagua R & Royuela M 2008 OSM, LIF, its receptors, and its relationship with the malignance in human breast carcinoma (in situ and in infiltrative). Cancer Investigation 26 222–229. doi:10.1080/07357900701638491.
Grant SL & Begley CG 1999 The oncostatin M signalling pathway: reversing the neoplastic phenotype? Molecular Medicine Today 5 406–412. doi:10.1016/S1357-4310(99)01540-3.
Grant SL, Hammacher A, Douglas AM, Goss GA, Mansfield RK, Heath JK & Begley CG 2002 An unexpected biochemical and functional interaction between gp130 and the EGF receptor family in breast cancer cells. Oncogene 21 460–474. doi:10.1038/sj.onc.1205100.
Guttilla IK, Phoenix KN, Hong X, Tirnauer JS, Claffey KP & White BA 2012 Prolonged mammosphere culture of MCF-7 cells induces an EMT and repression of the estrogen receptor by microRNAs. Breast Cancer Research and Treatment 132 75–85. doi:10.1007/s10549-011-1534-y.
Heinrich PC, Behrmann I, Haan S, Hermanns HM, Muller-Newen G & Schaper F 2003 Principles of interleukin (IL)-6-type cytokine signalling and its regulation. Biochemical Journal 374 1–20. doi:10.1042/BJ20030407.
Holzer RG, Ryan RE, Tommack M, Schlekeway E & Jorcyk CL 2004 Oncostatin M stimulates the detachment of a reservoir of invasive mammary carcinoma cells: role of cyclooxygenase-2. Clinical & Experimental Metastasis 21 167–176. doi:10.1023/B:CLIN.0000024760.02667.db.
Ikezoe T, Yang Y, Bandobashi K, Saito T, Takemoto S, Machida H, Togitani K, Koeffler HP & Taguchi H 2005 Oridonin, a diterpenoid purified from Rabdosia rubescens, inhibits the proliferation of cells from lymphoid malignancies in association with blockade of the NF-kappa B signal pathways. Molecular Cancer Therapeutics 4 578–586. doi:10.1158/1535-7163.MCT-04-0277.
Jones SA, Scheller J & Rose-John S 2011 Therapeutic strategies for the clinical blockade of IL-6/gp130 signaling. Journal of Clinical Investigation 121 3375–3383. doi:10.1172/JCI57158.
Jorcyk CL, Holzer RG & Ryan RE 2006 Oncostatin M induces cell detachment and enhances the metastatic capacity of T-47D human breast carcinoma cells. Cytokine 33 323–336. doi:10.1016/j.cyto.2006.03.004.
Kan CE, Cipriano R & Jackson MW 2011 c-MYC functions as a molecular switch to alter the response of human mammary epithelial cells to oncostatin M. Cancer Research 71 6930–6939. doi:10.1158/0008-5472.CAN-10-3860.
Kelly C, Jefferies C & Cryan SA 2011 Targeted liposomal drug delivery to monocytes and macrophages. Journal of Drug Delivery 2011 727241 doi:10.1155/2011/727241.
Kim MS, Louwagie J, Carvalho B, Terhaar Sive Droste JS, Park HL, Chae YK, Yamashita K, Liu J, Ostrow KL & Ling S et al. 2009 Promoter DNA methylation of oncostatin M receptor-β as a novel diagnostic and therapeutic marker in colon cancer. PLoS ONE 4 e6555 doi:10.1371/journal.pone.0006555.
Lopez-Tarruella S & Schiff R 2007 The dynamics of estrogen receptor status in breast cancer: re-shaping the paradigm. Clinical Cancer Research 13 6921–6925. doi:10.1158/1078-0432.CCR-07-1399.
Musgrove EA & Sutherland RL 2009 Biological determinants of endocrine resistance in breast cancer. Nature Reviews. Cancer 9 631–643. doi:10.1038/nrc2713.
Muss HB, Bunn JY, Crocker A, Plaut K, Koh J, Heintz N, Rincon M, Weaver DL, Tam D & Beatty B et al. 2007 Cyclin D-1, interleukin-6, HER-2/neu, transforming growth factor receptor-II and prediction of relapse in women with early stage, hormone receptor-positive breast cancer treated with tamoxifen. Breast Journal 13 337–345. doi:10.1111/j.1524-4741.2007.00440.x.
Naiditch H, Shurin MR & Shurin GV 2011 Targeting myeloid regulatory cells in cancer by chemotherapeutic agents. Immunologic Research 50 276–285. doi:10.1007/s12026-011-8213-2.
Needham LA, Davidson AH, Bawden LJ, Belfield A, Bone EA, Brotherton DH, Bryant S, Charlton MH, Clark VL & Davies SJ et al. 2011 Drug targeting to monocytes and macrophages using esterase-sensitive chemical motifs. Journal of Pharmacology and Experimental Therapeutics 339 132–142. doi:10.1124/jpet.111.183640.
Oh AS, Lorant LA, Holloway JN, Miller DL, Kern FG & El-Ashry D 2001 Hyperactivation of MAPK induces loss of ERα expression in breast cancer cells. Molecular Endocrinology 15 1344–1359. doi:10.1210/me.15.8.1344.
Osherov N & Levitzki A 1994 Epidermal-growth-factor-dependent activation of the src-family kinases. European Journal of Biochemistry 225 1047–1053. doi:10.1111/j.1432-1033.1994.1047b.x.
Prat A, Parker JS, Karginova O, Fan C, Livasy C, Herschkowitz JI, He X & Perou CM 2010 Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer. Breast Cancer Research 12 R68 doi:10.1186/bcr2635.
Queen MM, Ryan RE, Holzer RG, Keller-Peck CR & Jorcyk CL 2005 Breast cancer cells stimulate neutrophils to produce oncostatin M: potential implications for tumor progression. Cancer Research 65 8896–8904. doi:10.1158/0008-5472.CAN-05-1734.
Reid G, Denger S, Kos M & Gannon F 2002 Human estrogen receptor-α: regulation by synthesis, modification and degradation. Cellular and Molecular Life Sciences 59 821–831. doi:10.1007/s00018-002-8470-2.
Rody A, Holtrich U, Pusztai L, Liedtke C, Gaetje R, Ruckhaeberle E, Solbach C, Hanker L, Ahr A & Metzler D et al. 2009 T-cell metagene predicts a favorable prognosis in estrogen receptor-negative and HER2-positive breast cancers. Breast Cancer Research 11 R15 doi:10.1186/bcr2234.
Rubio MF, Werbajh S, Cafferata EG, Quaglino A, Colo GP, Nojek IM, Kordon EC, Nahmod VE & Costas MA 2006 TNF-α enhances estrogen-induced cell proliferation of estrogen-dependent breast tumor cells through a complex containing nuclear factor-kappa B. Oncogene 25 1367–1377. doi:10.1038/sj.onc.1209176.
Skliris GP, Hube F, Gheorghiu I, Mutawe MM, Penner C, Watson PH, Murphy LC, Leygue E & Myal Y 2008 Expression of small breast epithelial mucin (SBEM) protein in tissue microarrays (TMAs) of primary invasive breast cancers. Histopathology 52 355–369. doi:10.1111/j.1365-2559.2007.02955.x.
Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M & Jeffrey SS et al. 2001 Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. PNAS 98 10869–10874. doi:10.1073/pnas.191367098.
Sotiriou C & Pusztai L 2009 Gene-expression signatures in breast cancer. New England Journal of Medicine 360 790–800. doi:10.1056/NEJMra0801289.
Speirs V, Kerin MJ, Walton DS, Newton CJ, Desai SB & Atkin SL 2000 Direct activation of oestrogen receptor-α by interleukin-6 in primary cultures of breast cancer epithelial cells. British Journal of Cancer 82 1312–1316. doi:10.1054/bjoc.1999.1097.
Stearns V, Zhou Q & Davidson NE 2007 Epigenetic regulation as a new target for breast cancer therapy. Cancer Investigation 25 659–665. doi:10.1080/07357900701719234.
Storci G, Sansone P, Mari S, D'Uva G, Tavolari S, Guarnieri T, Taffurelli M, Ceccarelli C, Santini D & Chieco P et al. 2010 TNFα up-regulates SLUG via the NF-kappaB/HIF1α axis, which imparts breast cancer cells with a stem cell-like phenotype. Journal of Cellular Physiology 225 682–691. doi:10.1002/jcp.22264.
Stossi F, Madak-Erdogan Z & Katzenellenbogen BS 2011 Macrophage-elicited loss of estrogen receptor-α in breast cancer cells via involvement of MAPK and c-Jun at the ESR1 genomic locus. Oncogene doi:10.1038/onc.2011.370.
Szczepankiewicz BG, Kosogof C, Nelson LT, Liu G, Liu B, Zhao H, Serby MD, Xin Z, Liu M & Gum RJ et al. 2006 Aminopyridine-based c-Jun N-terminal kinase inhibitors with cellular activity and minimal cross-kinase activity. Journal of Medicinal Chemistry 49 3563–3580. doi:10.1021/jm060199b.
Tanaka M & Miyajima A 2003 Oncostatin M, a multifunctional cytokine. Reviews of Physiology, Biochemistry and Pharmacology 149 39–52. doi:10.1007/s10254-003-0013-1.
Teschendorff AE, Journee M, Absil PA, Sepulchre R & Caldas C 2007a Elucidating the altered transcriptional programs in breast cancer using independent component analysis. PLoS Computational Biology 3 e161 doi:10.1371/journal.pcbi.0030161.
Teschendorff AE, Miremadi A, Pinder SE, Ellis IO & Caldas C 2007b An immune response gene expression module identifies a good prognosis subtype in estrogen receptor negative breast cancer. Genome Biology 8 R157 doi:10.1186/gb-2007-8-8-r157.
Thompson JE, Cubbon RM, Cummings RT, Wicker LS, Frankshun R, Cunningham BR, Cameron PM, Meinke PT, Liverton N & Weng Y et al. 2002 Photochemical preparation of a pyridone containing tetracycle: a Jak protein kinase inhibitor. Bioorganic & Medicinal Chemistry Letters 12 1219–1223. doi:10.1016/S0960-894X(02)00106-3.
Underhill-Day N & Heath JK 2006 Oncostatin M (OSM) cytostasis of breast tumor cells: characterization of an OSM receptor β-specific kernel. Cancer Research 66 10891–10901. doi:10.1158/0008-5472.CAN-06-1766.
Vlahos CJ, Matter WF, Hui KY & Brown RF 1994 A specific inhibitor of phosphatidylinositol 3-kinase, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002). Journal of Biological Chemistry 269 5241–5248.
Vogt A, Qian Y, McGuire TF, Hamilton AD & Sebti SM 1996 Protein geranylgeranylation, not farnesylation, is required for the G1 to S phase transition in mouse fibroblasts. Oncogene 13 1991–1999.
Wang J, Barnes RO, West NR, Olson M, Chu JE & Watson PH 2008 Jab1 is a target of EGFR signaling in ERα-negative breast cancer. Breast Cancer Research 10 R51 doi:10.1186/bcr2105.
West NR & Watson PH 2010 S100A7 (psoriasin) is induced by the proinflammatory cytokines oncostatin-M and interleukin-6 in human breast cancer. Oncogene 29 2083–2092. doi:10.1038/onc.2009.488.
Whyte J, Bergin O, Bianchi A, McNally S & Martin F 2009 Key signalling nodes in mammary gland development and cancer. Mitogen-activated protein kinase signalling in experimental models of breast cancer progression and in mammary gland development. Breast Cancer Research 11 209 doi:10.1186/bcr2361.
Zhang F, Li C, Halfter H & Liu J 2003 Delineating an oncostatin M-activated STAT3 signaling pathway that coordinates the expression of genes involved in cell cycle regulation and extracellular matrix deposition of MCF-7 cells. Oncogene 22 894–905. doi:10.1038/sj.onc.1206158.