The ninth ENBDC Weggis meeting: growth and in-depth characterisation of normal and neoplastic breast cells.

Mammary gland biologists gathered for the ninth annual workshop of the European Network for Breast Development and Cancer (ENBDC) at Weggis on the shores of Lake Lucerne in March 2017. The main themes were oestrogen receptor alpha signalling, new techniques for mammary cell culture, CRISPR screening and proteogenomics.

Patient-derived xenograft (PDX) models in basic and translational breast cancer research.

Patient-derived xenograft (PDX) models of a growing spectrum of cancers are rapidly supplanting long-established traditional cell lines as preferred models for conducting basic and translational preclinical research. In breast cancer, to complement the now curated collection of approximately 45 long-established human breast cancer cell lines, a newly formed consortium of academic laboratories, currently from Europe, Australia, and North America, herein summarizes data on over 500 stably transplantable PDX models representing all three clinical subtypes of breast cancer (ER+, HER2+, and "Triple-negative" (TNBC)). Many of these models are well-characterized with respect to genomic, transcriptomic, and proteomic features, metastatic behavior, and treatment response to a variety of standard-of-care and experimental therapeutics. These stably transplantable PDX lines are generally available for dissemination to laboratories conducting translational research, and contact information for each collection is provided. This review summarizes current experiences related to PDX generation across participating groups, efforts to develop data standards for annotation and dissemination of patient clinical information that does not compromise patient privacy, efforts to develop complementary data standards for annotation of PDX characteristics and biology, and progress toward "credentialing" of PDX models as surrogates to represent individual patients for use in preclinical and co-clinical translational research. In addition, this review highlights important unresolved questions, as well as current limitations, that have hampered more efficient generation of PDX lines and more rapid adoption of PDX use in translational breast cancer research.

Erratum to: Humanization of the mouse mammary gland by replacement of the luminal layer with genetically engineered preneoplastic human cells.

Due to a typesetting error, the labelling was changed and the figures in this article [1] were presented in the order 2, 4, 10, 6, 1, 3, 5, 7, 8, 9, 11, 12, 13, 14 and the supplementary figure links were inverted. The revised version has the figures in the correct order.

Models incorporating chromatin modification data identify functionally important p53 binding sites.

Genome-wide prediction of transcription factor binding sites is notoriously difficult. We have developed and applied a logistic regression approach for prediction of binding sites for the p53 transcription factor that incorporates sequence information and chromatin modification data. We tested this by comparison of predicted sites with known binding sites defined by chromatin immunoprecipitation (ChIP), by the location of predictions relative to genes, by the function of nearby genes and by analysis of gene expression data after p53 activation. We compared the predictions made by our novel model with predictions based only on matches to a sequence position weight matrix (PWM). In whole genome assays, the fraction of known sites identified by the two models was similar, suggesting that there was little to be gained from including chromatin modification data. In contrast, there were highly significant and biologically relevant differences between the two models in the location of the predicted binding sites relative to genes, in the function of nearby genes and in the responsiveness of nearby genes to p53 activation. We propose that these contradictory results can be explained by PWM and ChIP data reflecting primarily biophysical properties of protein-DNA interactions, whereas chromatin modification data capture biologically important functional information.

Humanization of the mouse mammary gland by replacement of the luminal layer with genetically engineered preneoplastic human cells.

INTRODUCTION: The cell of origin for estrogen receptor α-positive (ERα+) breast cancer is probably a luminal stem cell in the terminal duct lobular units. To model these cells, we have used the murine myoepithelial layer in the mouse mammary ducts as a scaffold upon which to build a human luminal layer. To prevent squamous metaplasia, a common artifact in genetically-engineered breast cancer models, we sought to limit activation of the epidermal growth factor receptor (EGFR) during in vitro cell culture before grafting the cells. METHODS: Human reduction mammoplasty cells were grown in vitro in WIT medium. Epidermal growth factor in the medium was replaced with amphiregulin and neuregulin to decrease activation of EGFR and increase activation of EGFR homologs 3 and 4 (ERBB3 and ERBB4). Lentiviral vectors were used to express oncogenic transgenes and fluorescent proteins. Human mammary epithelial cells were mixed with irradiated mouse fibroblasts and Matrigel, then injected through the nipple into the mammary ducts of immunodeficient mice. Engrafted cells were visualized by stereomicroscopy for fluorescent proteins and characterized by histology and immunohistochemistry. RESULTS: Growth of normal mammary epithelial cells in conditions favoring ERBB3/4 signaling prevented squamous metaplasia in vitro. Normal human cells were quickly lost after intraductal injection, but cells infected with lentiviruses expressing CCND1, MYC, TERT, BMI1 and a short-hairpin RNA targeting TP53 were able to engraft and progressively replace the luminal layer in the mouse mammary ducts, resulting in the formation of an extensive network of humanized ducts. Despite expressing multiple oncogenes, the human cells formed a morphologically normal luminal layer. Expression of a single additional oncogene, PIK3CA-H1047R, converted the cells into invasive cancer cells. The resulting tumors were ERα+, Ki67+ luminal B adenocarcinomas that were resistant to treatment with fulvestrant. CONCLUSIONS: Injection of preneoplastic human mammary epithelial cells into the mammary ducts of immunodeficient mice leads to replacement of the murine luminal layer with morphologically normal human cells. Genetic manipulation of the injected cells makes it possible to study defined steps in the transformation of human mammary epithelial cells in a more physiological environment than has hitherto been possible.

New insights into the role of androgen and oestrogen receptors in molecular apocrine breast tumours.

Two recent studies on a rare androgen-dependent form of breast cancer have shed light on the biology of luminal tumours and reinforced the view that interfering with androgen signalling may have a place in the therapy of some forms of breast cancer.

Validation of gene signatures that predict the response of breast cancer to neoadjuvant chemotherapy: a substudy of the EORTC 10994/BIG 00-01 clinical trial.

BACKGROUND: We have previously described gene-expression signatures that predict growth inhibitory and cytotoxic effects of common chemotherapeutic drugs in vitro. The aim of this study was to confirm the validity of these gene-expression signatures in a large series of patients with oestrogen-receptor-negative breast tumours who were treated in a phase III neoadjuvant clinical trial. METHODS: This trial compares a non-taxane regimen (fluorouracil, epirubicin, and cyclophosphamide [FEC] for six cycles) with a taxane regimen (docetaxel for three cycles followed by epirubicin plus docetaxel [TET] for three cycles) in women with oestrogen-receptor-negative breast cancer. The primary endpoint of the study is the difference in progression-free survival based on TP53 status and will be reported later. Predicting response with gene signatures was a planned secondary endpoint of the trial and is reported here. Pathological complete response, defined as complete disappearance of the tumour with no more than a few scattered tumour cells detected by the pathologist in the resection specimen, was used to assess chemosensitivity. RNA was prepared from sections of frozen biopsies taken at diagnosis and hybridised to Affymetrix X3P microarrays. In-vitro single-agent drug sensitivity signatures were combined to obtain FEC and TET regimen-specific signatures. This study is registered on the clinical trials site of the US National Cancer Institute website http://www.clinicaltrials.gov/ct/show/NCT00017095. FINDINGS: Of 212 patients with oestrogen-receptor-negative tumours assessed, 87 patients were excluded. 125 oestrogen-receptor-negative tumours (55 that showed pathological complete responses) were tested: 66 in the FEC group (28 that showed pathological complete responses) and 59 in the TET group (27 that showed pathological complete responses). The regimen-specific signatures significantly predicted pathological complete response in patients treated with the appropriate regimen (p<0.0001). The FEC predictor had a sensitivity of 96% (27 of 28 patients [95% CI 82-99]), specificity of 66% (25 of 38 patients [50-79]), positive predictive value (PPV) of 68% (27 of 40 patients [52-80]), and negative predictive value (NPV) of 96% (25 of 26 patients [81-99]). The TET predictor had a sensitivity of 93% (25 of 27 patients [77-98]), specificity 69% (22 of 32 patients [51-82]), PPV of 71% (25 of 35 patients [55-84]), and NPV of 92% (22 of 24 patients [74-98]). Analysis of tumour size, grade, nodal status, age, and regimen-specific signatures showed that the genomic signatures were the only independent variables predicting pathological complete response at p<0.01. Selection of patients with these signatures would increase the proportion of patients with pathological complete responses from 44% to around 70% in the patients studied here. INTERPRETATION: We have validated the use of regimen-specific drug sensitivity signatures in the context of a multicentre randomised trial. The high NPV of both signatures may allow early selection of patients with breast cancer who should be considered for trials with new drugs.

An oestrogen-dependent model of breast cancer created by transformation of normal human mammary epithelial cells.

INTRODUCTION: About 70% of breast cancers express oestrogen receptor alpha (ESR1/ERalpha) and are oestrogen-dependent for growth. In contrast with the highly proliferative nature of ERalpha-positive tumour cells, ERalpha-positive cells in normal breast tissue rarely proliferate. Because ERalpha expression is rapidly lost when normal human mammary epithelial cells (HMECs) are grown in vitro, breast cancer models derived from HMECs are ERalpha-negative. Currently only tumour cell lines are available to model ERalpha-positive disease. To create an ERalpha-positive breast cancer model, we have forced normal HMECs derived from reduction mammoplasty tissue to express ERalpha in combination with other relevant breast cancer genes. METHODS: Candidate genes were selected based on breast cancer microarray data and cloned into lentiviral vectors. Primary HMECs prepared from reduction mammoplasty tissue were infected with lentiviral particles. Infected HMECs were characterised by Western blotting, immunofluorescence microscopy, microarray analysis, growth curves, karyotyping and SNP chip analysis. The tumorigenicity of the modified HMECs was tested after orthotopic injection into the inguinal mammary glands of NOD/SCID mice. Cells were marked with a fluorescent protein to allow visualisation in the fat pad. The growth of the graft was analysed by fluorescence microscopy of the mammary glands and pathological analysis of stained tissue sections. Oestrogen dependence of tumour growth was assessed by treatment with the oestrogen antagonist fulvestrant. RESULTS: Microarray analysis of ERalpha-positive tumours reveals that they commonly overexpress the Polycomb-group gene BMI1. Lentiviral transduction with ERalpha, BMI1, TERT and MYC allows primary HMECs to be expanded in vitro in an oestrogen-dependent manner. Orthotopic xenografting of these cells into the mammary glands of NOD/SCID mice results in the formation of ERalpha-positive tumours that metastasise to multiple organs. The cells remain wild type for TP53, diploid and genetically stable. In vivo tumour growth and in vitro proliferation of cells explanted from tumours are dependent on oestrogen. CONCLUSION: We have created a genetically defined model of ERalpha-positive human breast cancer based on normal HMECs that has the potential to model human oestrogen-dependent breast cancer in a mouse and enables the study of mechanisms involved in tumorigenesis and metastasis.

RAD001 (everolimus) improves the efficacy of replicating adenoviruses that target colon cancer.

Selectively replicating adenoviruses have the potential to cure cancer but have shown little efficacy in clinical trials. We have tested the ability of the mTOR kinase inhibitor RAD001 (everolimus) to enhance the response of xenografts to an oncolytic adenovirus. The virus has Tcf sites inserted in the early viral promoters and replicates selectively in cells with activation of the Wnt signaling pathway. To enhance tumor cell infection, an integrin targeting peptide (CDCRGDCFC) was inserted into the fiber gene of the virus. RAD001 combines three useful properties: it inhibits tumor cell growth directly, blocks angiogenesis, and suppresses the immune response. RAD001 does not block viral protein expression, DNA replication, or cytopathic effect in tumor cells in vitro. After 6 weeks of daily RAD001 treatment, ongoing viral DNA replication could be detected in tumor xenografts, showing that RAD001 does not inhibit virus replication in vivo. I.v. injection of virus alone produced a small delay in xenograft growth, whereas combination therapy substantially prolonged the survival of the mice. We suggest that collapsing the tumor vasculature after the initial infection traps the virus and facilitates local spread within the tumor. Unlike conventional drugs, which require continued access to the tumor through the vascular system, oncolytic viruses are in principle less sensitive to late reductions in perfusion because they are produced locally within the tumor.

Fusion of the BCL9 HD2 domain to E1A increases the cytopathic effect of an oncolytic adenovirus that targets colon cancer cells.

BACKGROUND: The Wnt signaling pathway is activated by mutations in the APC and beta-catenin genes in many types of human cancer. beta-catenin is stabilized by these mutations and activates transcription in part by acting as a bridge between Tcf/LEF proteins and the HD2 domain of the BCL9 coactivator. We have previously described oncolytic adenoviruses with binding sites for Tcf/LEF transcription factors inserted into the early viral promoters. These viruses replicate selectively in cells with activation of the Wnt pathway. To increase the activity of these viruses we have fused the viral transactivator E1A to the BCL9 HD2 domain. METHODS: Luciferase assays, co-immunoprecipitation and Western blotting, immunofluorescent cell staining and cytopathic effect assays were used to characterize the E1A-HD2 fusion protein and virus in vitro. Growth curves of subcutaneous SW620 colon cancer xenografts were used to characterize the virus in vivo. RESULTS: The E1A-HD2 fusion protein binds to beta-catenin in vivo and activates a Tcf-regulated luciferase reporter better than wild-type E1A in cells with activated Wnt signaling. Expression of the E1A-HD2 protein promotes nuclear import of beta-catenin, mediated by the strong nuclear localization signal in E1A. Tcf-regulated viruses expressing the fusion protein show increased expression of viral proteins and a five-fold increase in cytopathic effect (CPE) in colorectal cancer cell lines. There was no change in viral protein expression or CPE in HeLa cells, indicating that E1A-HD2 viruses retain selectivity for cells with activation of the Wnt signaling pathway. Despite increasing the cytopathic effect of the virus in vitro, fusion of the HD2 domain to E1A did not increase the burst size of the virus in vitro or the anti-tumor effect of the virus in an SW620 xenograft model in vivo. CONCLUSION: Despite an increase in the nuclear pool of beta-catenin, the effects on viral activity in colon cancer cells were small, suggesting that factors acting downstream of beta-catenin are limiting for viral replication and toxicity in these cells. The approach of fusing E1A to a protein domain implicated in oncogenic signaling could be used to selectively increase the activity of oncolytic viruses targeting several other pathways defective in cancer.

Cluster Analysis of p53 Binding Site Sequences Reveals Subsets with Different Functions.

p53 is an important regulator of cell cycle arrest, senescence, apoptosis and metabolism, and is frequently mutated in tumors. It functions as a tetramer, where each component dimer binds to a decameric DNA region known as a response element. We identify p53 binding site subtypes and examine the functional and evolutionary properties of these subtypes. We start with over 1700 known binding sites and, with no prior labeling, identify two sets of response elements by unsupervised clustering. When combined, they give rise to three types of p53 binding sites. We find that probabilistic and alignment-based assessments of cross-species conservation show no strong evidence of differential conservation between types of binding sites. In contrast, functional analysis of the genes most proximal to the binding sites provides strong bioinformatic evidence of functional differentiation between the three types of binding sites. Our results are consistent with recent structural data identifying two conformations of the L1 loop in the DNA binding domain, suggesting that they reflect biologically meaningful groups imposed by the p53 protein structure.

Promoter-specific p53-dependent histone acetylation following DNA damage.

We have used chromatin immunoprecipitation (ChIP) to measure p53-dependent histone acetylation at the p21, MDM2, and PUMA promoters. The pattern of histone acetylation was different at each promoter. H3 and H4 acetylation increased at both the p21 and PUMA promoters in response to p53 activation, whereas there was only a minimal increase in H4 acetylation and no increase in H3 acetylation at the MDM2 promoter. The high p53 occupancy of the p21, MDM2 and PUMA promoters has been attributed to the presence of two p53 binding sites in these promoters, but mutation of the p53 binding sites in integrated p21 promoter constructs showed that the two sites in the p21 promoter do not cooperate to stabilize p53 binding. Despite 10-fold higher p53 binding to the proximal than the distal site in the p21 promoter, both sites showed similar patterns of H3 and H4 acetylation. Mutation of the binding sites showed that acetylation of the proximal, low-affinity site requires p53 binding to that site but not to the distal, high-affinity site. Since low-affinity p53 binding sites can confer strong acetylation, the DNA binding affinity in vitro is an unreliable guide to the likely importance of p53 in regulating candidate target genes in vivo.

Evidence of frequent recombination among human adenoviruses.

Genome stability is a prerequisite for the production and use of adenoviruses for therapy of genetic diseases and cancer. To test the premise that the adenoviral genome is stable, the phylogenetic relationships of 16 adenovirus C (AdC) field isolates were studied in four genome regions: hexon, fiber, polymerase and E1A. The phylogenetic relationships in the fiber gene concurred with those in the hexon region. In contrast, the non-structural regions had marks of frequent recombination, to the point that an isolate of one serotype could contain non-structural proteins that were identical to the genes from a different serotype. Our results suggest that recombination among circulating adenoviruses is very frequent and plays an important role in shaping the phylogenetic relationships of adenovirus genomes. Analysis of the available complete genome sequences of AdB, AdC and AdD species showed that recombination shuffles genome fragments within a species, but not between species. One of the AdC field isolates possessed the fiber gene of AdC type 6, but a hexon gene that was distinct from all AdC serotypes. This strain could not be typed unambiguously in a neutralization test and might represent a novel serotype of AdC. Comparison of the right end (nt 18838-33452) of this isolate with that of the ATCC Ad6 strain showed clear evidence of multiple recombination events.

Expression of heterologous genes in oncolytic adenoviruses using picornaviral 2A sequences that trigger ribosome skipping.

Insertion of picornaviral 2A sequences into mRNAs causes ribosomes to skip formation of a peptide bond at the junction of the 2A and downstream sequences, leading to the production of two proteins from a single open reading frame. Adenoviral protein IX is a minor capsid protein that has been used to display foreign peptides on the surface of the capsid. We have used 2A sequences from the foot-and-mouth disease virus (FMDV) and porcine teschovirus 1 (PTV-1) to express protein IX (pIX) and green fluorescent protein (GFP) from pIX-2A-GFP fusion genes in an oncolytic virus derived from human adenovirus 5. GFP was efficiently expressed by constructs containing either 2A sequence. Peptide bond skipping was more efficient with the 58 aa FMDV sequence than with the 22 aa PTV-1 2A sequence, but the virus with the FMDV 2A sequence showed a reduction in plaque size, cytopathic effect, viral burst size and capsid stability. We conclude that ribosome skipping induced by 2A sequences is an effective strategy to express heterologous genes in adenoviruses; however, careful selection or optimization of the 2A sequence may be required if protein IX is used as the fusion partner.

Determinants of interferon-stimulated gene induction by RNAi vectors.

RNA interference is widely used to silence gene expression in mammalian cells. We recently reported that an shRNA expressed from the H1 promoter in a lentiviral vector could induce the expression of a large group of interferon-stimulated genes (ISGs). This response was unrelated to silencing of the gene targeted by the shRNA MORF4L1. In parallel, we constructed lentiviral vectors expressing shRNA from the U6 promoter and found that these too could induce expression of OAS1, a classic interferon target gene. The U6 vectors give a higher frequency of ISG induction than comparable lentiviral H1 vectors, suggesting that there might be a fundamental flaw in the vector design. We have characterized the U6 vectors in detail and report here that ISG induction is a consequence of the presence of an AA di-nucleotide near the transcription start site. A single nucleotide deletion in the siRNA sequence abolished OAS1 induction, suggesting that the mechanism underlying the response uses a sensor that can detect 19 bp RNA duplexes but not 14 bp duplexes. Adenoviral VA RNA I, which inhibits dsRNA-dependent protein kinase (PKR), was tested as a fusion partner to express shRNA on the grounds that it might prevent nonspecific off-target effects. Fusion of VA RNA I to a lamin shRNA was moderately effective in silencing lamin expression, but gave strong OAS1 induction by an shRNA that does not induce OAS1 when expressed from the U6 or H1 promoters. To avoid interferon induction by U6 vectors, we recommend preserving the wild-type sequence around the transcription start site, in particular a C/G sequence at positions -1/+1, and we describe a simple cloning strategy using the Gateway recombination system that facilitates this task.

Regulation of p53 stability and function in HCT116 colon cancer cells.

We have used a lentiviral vector to stably express p53 at a physiological level in p53 knockout HCT116 cells. Cells transduced with wild type p53 responded to genotoxic stress by stabilizing p53 and expressing p53 target genes. The reconstituted cells underwent G(1) arrest or apoptosis appropriately depending on the type of stress, albeit less efficiently than parental wild type cells. Compared with cells expressing exogenous wild type p53, the apoptotic response to 5-fluorouracil (5FU) was >50% reduced in cells expressing S15A or S20A mutant p53, and even more reduced by combined mutation of serines 6, 9, 15, 20, 33, and 37 (N6A). Among a panel of p53 target genes tested by quantitative PCR, the gene showing the largest defect in induction by 5FU was BBC3 (PUMA), which was induced 4-fold by wild type p53 and 2-fold by the N6A mutant. Mutation of N-terminal phosphorylation sites did not prevent p53 stabilization by doxorubicin or 5FU. MDM2 silencing by RNA interference activated p53 target gene expression in normal fibroblasts but not in HCT116 cells, and exogenous p53 could be stabilized in HCT116 knockout cells despite combined mutation of p53 phosphorylation sites and silencing of MDM2 expression. The MDM2 feedback loop is thus defective, and other mechanisms must exist to regulate p53 stability and function in this widely used tumor cell line.