Identification of novel neuroendocrine-specific tumour genes

Neuroendocrine gene subsets are uniquely dysregulated in prostate adenocarcinoma

Prostate cancer has heterogeneous growth patterns, and its prognosis is the poorest when it progresses to a neuroendocrine phenotype. Using bioinformatic analysis, we evaluated RNA expression of neuroendocrine genes in a panel of five different cancer types: prostate adenocarcinoma, breast cancer, kidney chromophobe, kidney renal clear cell carcinoma and kidney renal papillary cell carcinoma. Our results show that specific neuroendocrine genes are significantly dysregulated in these tumors, suggesting that they play an active role in cancer progression. Among others, synaptophysin (SYP), a conventional neuroendocrine marker, is upregulated in prostate adenocarcinoma (PRAD) and breast cancer (BRCA). Our analysis shows that SYP is enriched in small extracellular vesicles (sEVs) derived from plasma of PRAD patients, but it is absent in sEVs derived from plasma of healthy donors. Similarly, classical sEV markers are enriched in sEVs derived from plasma of prostate cancer patients, but weakly detectable in sEVs derived from plasma of healthy donors. Overall, our results pave the way to explore new strategies to diagnose these diseases based on the neuroendocrine gene expression in patient tumors or plasma sEVs.

CDHu40: a novel marker gene set of neuroendocrine prostate cancer

Prostate cancer (PCa) is the most prevalent cancer affecting American men. Castration-resistant prostate cancer (CRPC) can emerge during hormone therapy for PCa, manifesting with elevated serum prostate-specific antigen levels, continued disease progression, and/or metastasis to the new sites, resulting in a poor prognosis. A subset of CRPC patients shows a neuroendocrine (NE) phenotype, signifying reduced or no reliance on androgen receptor signaling and a particularly unfavorable prognosis. In this study, we incorporated computational approaches based on both gene expression profiles and protein-protein interaction networks. We identified 500 potential marker genes, which are significantly enriched in cell cycle and neuronal processes. The top 40 candidates, collectively named CDHu40, demonstrated superior performance in distinguishing NE PCa (NEPC) and non-NEPC samples based on gene expression profiles. CDHu40 outperformed most of the other published marker sets, excelling particularly at the prognostic level. Notably, some marker genes in CDHu40, absent in the other marker sets, have been reported to be associated with NEPC in the literature, such as DDC, FOLH1, BEX1, MAST1, and CACNA1A. Importantly, elevated CDHu40 scores derived from our predictive model showed a robust correlation with unfavorable survival outcomes in patients, indicating the potential of the CDHu40 score as a promising indicator for predicting the survival prognosis of those patients with the NE phenotype. Motif enrichment analysis on the top candidates suggests that REST and E2F6 may serve as key regulators in the NEPC progression.

Applied models and molecular characteristics of small cell lung cancer

Small cell lung cancer (SCLC) is a highly aggressive type of cancer frequently diagnosed with metastatic spread, rendering it surgically unresectable for the majority of patients. Although initial responses to platinum-based therapies are often observed, SCLC invariably relapses within months, frequently developing drug-resistance ultimately contributing to short overall survival rates. Recently, SCLC research aimed to elucidate the dynamic changes in the genetic and epigenetic landscape. These have revealed distinct subtypes of SCLC, each characterized by unique molecular signatures. The recent understanding of the molecular heterogeneity of SCLC has opened up potential avenues for precision medicine, enabling the development of targeted therapeutic strategies. In this review, we delve into the applied models and computational approaches that have been instrumental in the identification of promising drug candidates. We also explore the emerging molecular diagnostic tools that hold the potential to transform clinical practice and patient care.

CDHu40: a novel marker gene set of neuroendocrine prostate cancer (NEPC)

Prostate cancer (PCa) is the most prevalent cancer affecting American men. Castration-resistant prostate cancer (CRPC) can emerge during hormone therapy for PCa, manifesting with elevated serum prostate-specific antigen (PSA) levels, continued disease progression, and/or metastasis to the new sites, resulting in a poor prognosis. A subset of CRPC patients shows a neuroendocrine (NE) phenotype, signifying reduced or no reliance on androgen receptor (AR) signaling and a particularly unfavorable prognosis. In this study, we incorporated computational approaches based on both gene expression profiles and protein-protein interaction (PPI) networks. We identified 500 potential marker genes, which are significantly enriched in cell cycle and neuronal processes. The top 40 candidates, collectively named as CDHu40, demonstrated superior performance in distinguishing NE prostate cancer (NEPC) and non-NEPC samples based on gene expression profiles compared to other published marker sets. Notably, some novel marker genes in CDHu40, absent in the other marker sets, have been reported to be associated with NEPC in the literature, such as DDC, FOLH1, BEX1, MAST1, and CACNA1A. Importantly, elevated CDHu40 scores derived from our predictive model showed a robust correlation with unfavorable survival outcomes in patients, indicating the potential of the CDHu40 score as a promising indicator for predicting the survival prognosis of those patients with the NE phenotype. Motif enrichment analysis on the top candidates suggests that REST and E2F6 may serve as key regulators in the NEPC progression.

Significance

our study integrates gene expression variances in multiple NEPC studies and protein-protein interaction network to pinpoint a specific set of NEPC maker genes namely CDHu40. These genes and scores based on their gene expression levels effectively distinguish NEPC samples and underscore the clinical prognostic significance and potential mechanism.

BEX1 mediates sorafenib resistance in hepatocellular carcinoma by regulating AKT signaling

Sorafenib is the first-line therapy for advanced hepatocellular carcinoma (HCC). However, acquired tolerance after sorafenib treatment significantly limits its therapeutic efficacy, and the mechanisms underlying resistance remains poorly characterized. In this study, we identified BEX1 as key mediator of sorafenib resistance in HCC. We found that BEX1 expression was significantly reduced in sorafenib-resistant HCC cells and xenograft models, moreover, BEX1 expression in HCC tissues was down-regulated compared with that normal liver tissues in The Cancer Genome Atlas (TCGA) database, K-M analysis demonstrated that reduced BEX1 expression was correlated with poor clinical prognosis in HCC patients. Loss- and gain-of-function studies showed that BEX1 regulates the cell-killing ability of sorafenib. Further studies revealed that BEX1 renders HCC cells sensitive to sorafenib via induction of apoptosis and negatively regulates the phosphorylation of Akt. In summary, our study uncover BEX1 may serve as a promising predictive biomarker for the prognosis of patients with HCC.

Comprehensive machine-learning survival framework develops a consensus model in large-scale multicenter cohorts for pancreatic cancer

As the most aggressive tumor, the outcome of pancreatic cancer (PACA) has not improved observably over the last decade. Anatomy-based TNM staging does not exactly identify treatment-sensitive patients, and an ideal biomarker is urgently needed for precision medicine. Based on expression files of 1280 patients from 10 multicenter cohorts, we screened 32 consensus prognostic genes. Ten machine-learning algorithms were transformed into 76 combinations, of which we selected the optimal algorithm to construct an artificial intelligence-derived prognostic signature (AIDPS) according to the average C-index in the nine testing cohorts. The results of the training cohort, nine testing cohorts, Meta-Cohort, and three external validation cohorts (290 patients) consistently indicated that AIDPS could accurately predict the prognosis of PACA. After incorporating several vital clinicopathological features and 86 published signatures, AIDPS exhibited robust and dramatically superior predictive capability. Moreover, in other prevalent digestive system tumors, the nine-gene AIDPS could still accurately stratify the prognosis. Of note, our AIDPS had important clinical implications for PACA, and patients with low AIDPS owned a dismal prognosis, higher genomic alterations, and denser immune cell infiltrates as well as were more sensitive to immunotherapy. Meanwhile, the high AIDPS group possessed observably prolonged survival, and panobinostat may be a potential agent for patients with high AIDPS. Overall, our study provides an attractive tool to further guide the clinical management and individualized treatment of PACA.

A homozygous nonsense mutation in DCBLD2 is a candidate cause of developmental delay, dysmorphic features and restrictive cardiomyopathy

DCBLD2 encodes discodin, CUB and LCCL domain-containing protein 2, a type-I transmembrane receptor that is involved in intracellular receptor signalling pathways and the regulation of cell growth. In this report, we describe a 5-year-old female who presented severe clinical features, including restrictive cardiomyopathy, developmental delay, spasticity and dysmorphic features. Trio-whole-exome sequencing and segregation analysis were performed to identify the genetic cause of the disease within the family. A novel homozygous nonsense variant in the DCBLD2 gene (c.80G > A, p.W27*) was identified as the most likely cause of the patient's phenotype. This nonsense variant falls in the extracellular N-terminus of DCBLD2 and thus might affect proper protein function of the transmembrane receptor. A number of in vitro investigations were performed on the proband's skin fibroblasts compared to normal fibroblasts, which allowed a comprehensive assessment resulting in the functional characterization of the identified DCBLD2 nonsense variant in different cellular processes. Our data propose a significant association between the identified variant and the observed reduction in cell proliferation, cell cycle progression, intracellular ROS, and Ca2 + levels, which would likely explain the phenotypic presentation of the patient as associated with lethal restrictive cardiomyopathy.

DCBLD2 Affects the Development of Colorectal Cancer via EMT and Angiogenesis and Modulates 5-FU Drug Resistance

Background: DCBLD2 is highly expressed in various cancers, including colorectal cancer. DCBLD2 overexpression promotes tumor occurrence, development, and metastasis. However, DCBLD2 sensitivity to chemotherapy drugs and its mechanism on tumor development are unknown. Methods: DCBLD2 expression differences in cancer and normal tissues were obtained from GEO and TCGA databases. DCBLD2 influence on prognosis was also compared, and the database analysis results were verified via the analysis of clinical samples. GDSC database was used to analyze the effect of DCBLD2 expression difference on 5-FU drug sensitivity on tumor cells. CCK-8, clone formation, scratch, Transwell invasion and migration assays were used to assess DCBLD2 effects on the proliferation, metastasis, and 5-FU drug sensitivity on HCT116 and Caco-2 colorectal cancer cells. Angiogenesis and Matrigel plug assays were used to study the effect of DCBLD2 on angiogenesis. Q-RCR and Western Blot were used to analyze DCBLD2 impact on the EMT signaling pathway, and TAP-MS assay with Co-IP verification was used to identify the downstream target proteins binding to DCBLD2. Results: Both database and clinical sample validation results showed that the expression of DCBLD2 in colorectal cancer tissues was significantly higher than that in normal tissues, leading to poor prognosis of patients. GDSC database analysis showed that DCBLD2 overexpression caused tumor cell resistance to 5-FU. The results of in vitro and in vivo experiments showed that the inhibition of DCBLD2 reduced the proliferation, migration and invasion of colorectal cancer cells, inhibited the angiogenesis of endothelial cells, and enhanced the drug sensitivity to 5-FU. The results of q-RCR and Western Blot experiments showed that the inhibition of DCBLD2 can suppress the EMT signal. The results of TAP-MS assay showed that the proteins bound to DCBLD2 were enriched to the Focal adhesion pathway. The results of Co-IP assay show that DCBLD2 can combine with ITGB1, the key factor of Focal adhesion pathway. Conclusion: DCBLD2 may affect the development of colorectal cancer by regulating cell proliferation and motility, and modulate 5-FU resistance. Down-regulation of DCBLD2 can inhibit EMT signal and angiogenesis. DCBLD2 can combine with ITGB1, the key signal factor of the Focal adhesion pathway.

ESDN inhibits melanoma progression by blocking E-selectin expression in endothelial cells via STAT3

An interactive crosstalk between tumor and stroma cells is essential for metastatic melanoma progression. We evidenced that ESDN/DCBLD2/CLCP1 plays a crucial role in endothelial cells during the spread of melanoma. Precisely, increased extravasation and metastasis formation were revealed in ESDN-null mice injected with melanoma cells, even if the primary tumor growth, vessel permeability, and angiogenesis were not enhanced. Interestingly, improved adhesion of melanoma cells to ESDN-depleted endothelial cells was observed, due to the presence of higher levels of E-selectin transcripts/proteins in ESDN-defective cells. In accordance with these results, anticorrelation was observed between ESDN and E-selectin in human endothelial cells. Most importantly, our data revealed that cimetidine, an E-selectin inhibitor, was able to block cell adhesion, extravasation, and metastasis formation in ESDN-null mice, underlying a major role of ESDN in E-selectin transcription upregulation, which according to our data, may presumably be linked to STAT3. Based on our results, we propose a protective role for ESDN during the spread of melanoma and reveal its therapeutic potential.

DCBLD2 Mediates Epithelial-Mesenchymal Transition-Induced Metastasis by Cisplatin in Lung Adenocarcinoma

Growing evidence suggests that cisplatin and other chemotherapeutic agents promote tumor metastasis while inhibiting tumor growth, which is a critical issue for certain patients in clinical practices. However, the role of chemotherapeutics in promoting tumor metastasis and the molecular mechanism involved are unclear. Here, we investigated the roles of cisplatin in promoting tumor metastasis in lung adenocarcinoma (LUAD). We demonstrated that cisplatin promoted epithelial-mesenchymal transition (EMT), cell motility, and metastasis in vitro and in vivo. The bioinformatic analysis and molecular biology approaches also indicated that DCBLD2 (Discoidin, CUB and LCCL domain containing 2) is a key gene that mediates cisplatin-induced metastasis. DCBLD2 stabilizes β-catenin by phosphorylating GSK3β and transporting accumulated β-catenin to the nucleus to promote the expression of EMT-related transcriptional factors (TFs), ultimately resulting in tumor metastasis. We also identified that cisplatin enhanced DCBLD2 expression by phosphorylating ERK and hence the AP-1-driven transcription of DCBLD2. Furthermore, DCBLD2-specific siRNAs encapsulated by nanocarriers prominently inhibit cisplatin-induced metastasis in vivo. Therefore, DCBLD2 plays a key role in cisplatin-induced metastasis in LUAD and is a potential target for preventing chemotherapy-induced metastasis in vivo.

Bex1 significantly contributes to the proliferation and invasiveness of malignant tumor cells

Invasion has a significant role in cancer progression, including expansion to surrounding tissue and metastasis. Previously, we assessed the invasive ability of cancer cells using an easy-to-prepare double-layered collagen gel hemisphere (DL-CGH) method by which cancer cell invasion can be easily visualized. The present study examined multiple lung adenocarcinoma and malignant pleural mesothelioma (MPM) cell lines using the DL-CGH method and identified inherently invasive cell lines. Next, by comparing gene expression between invasive and non-invasive cells by cDNA microarray, the potential candidate gene brain-expressed x-linked protein 1 (Bex1) was identified to be involved in cancer invasion, as it was highly expressed in the invasive cell lines. Downregulation of Bex1 suppressed the invasion and proliferation of the invasive tumor cell lines. The findings of the present study suggested that Bex1 may promote metastasis in vivo and could be a potential oncogene and molecular therapeutic target in lung adenocarcinoma and MPM.

FYN and ABL Regulate the Interaction Networks of the DCBLD Receptor Family

The Discoidin, CUB, and LCCL domain-containing protein (DCBLD) family consists of two type-I transmembrane scaffolding receptors, DCBLD1 and DCBLD2, which play important roles in development and cancer. The nonreceptor tyrosine kinases FYN and ABL are known to drive phosphorylation of tyrosine residues in YXXP motifs within the intracellular domains of DCBLD family members, which leads to the recruitment of the Src homology 2 (SH2) domain of the adaptors CT10 regulator of kinase (CRK) and CRK-like (CRKL). We previously characterized the FYN- and ABL-driven phosphorylation of DCBLD family YXXP motifs. However, we have identified additional FYN- and ABL-dependent phosphorylation sites on DCBLD1 and DCBLD2. This suggests that beyond CRK and CRKL, additional DCBLD interactors may be regulated by FYN and ABL activity. Here, we report a quantitative proteomics approach in which we map the FYN- and ABL-regulated interactomes of DCBLD family members. We found FYN and ABL regulated the binding of several signaling molecules to DCBLD1 and DCBLD2, including members of the 14-3-3 family of adaptors. Biochemical investigation of the DCBLD2/14-3-3 interaction revealed ABL-induced binding of 14-3-3 family members directly to DCBLD2.

Identification of new candidate genes and signalling pathways associated with the development of neuroendocrine pancreatic tumours based on next generation sequencing data

Despite advances in classification, treatment, and imaging, neuroendocrine tumours remain a clinically complex subject. In this work, we studied the genetic profile of well-differentiated pancreatic neuroendocrine tumours (PanNETs) in a cohort of Caucasian patients and analysed the signalling pathways and candidate genes potentially associated with the development of this oncological disease. Twenty-four formalin-fixed paraffin-embedded (FFPE) samples of well-differentiated PanNETs were subjected to massive parallel sequencing using the targeted gene panel (409 genes) of the Illumina NextSeq 550 platform (San Diego, USA). In 24 patients, 119 variants were identified in 54 genes. The median mutation rate per patient was 5 (2.8-7). The detected genetic changes were dominated by missense mutations (67%) and nonsense mutations (29%). 18% of the mutations were activating, 35% of the variants led to a loss of function of the encoded protein, and 52% were not classified. Twenty-six variants were described as new. Functionally significant changes in the tertiary structure and activity of the protein molecules in an in silico assay were predicted for 5 new genetic variants. The 5 highest priority candidate genes were selected: CREB1, TCF12, PRKAR1A, BCL11A, and BUB1B. Genes carrying the identified mutations participate in signalling pathways known to be involved in PanNETs; in addition, 38% of the cases showed genetic changes in the regulation of the SMAD2/3 signalling pathway. Well-differentiated PanNETs in a Russian cohort demonstrate various molecular genetic features, including new genetic variations and potential driver genes. The highlighted molecular genetic changes in the SMAD2/3 signalling pathway suggest new prospects for targeted therapy.

STAT3 Inhibition for Gastroenteropancreatic Neuroendocrine Tumors: Potential for a New Therapeutic Target?

BACKGROUND

Gastroenteropancreatic neuroendocrine tumors (GEP-NETs) are highly vascular neoplasms treated similarly, irrespective of tumor location. The expression of pro-angiogenic factors (STAT3, VEGF, and HIF-1α) and their association with adverse pathologic factors and disease recurrence following resection remains unclear.

METHODS

All patients with non-metastatic GEP-NETs who underwent curative-intent resection from 2000 to 2013 were included. Immunohistochemistry was performed for pro-angiogenic factors, Ki-67 index, and CD31 using tissue microarrays made in triplicate by a pathologist blinded to other clinicopathologic variables. Primary outcome was a 3-year recurrence-free survival (3-yrRFS); secondary outcomes were correlation of pro-angiogenic factors with Ki-67 index, adverse pathologic factors, and CD31 expression, a marker of microvascular density.

RESULTS

Of 144 GEP-NETs resected, STAT3 expression was high in 12 (8%) and low in 132 (92%) pts. High STAT3 expression was associated with worse 3-yrRFS compared to low expression (55% vs 84%; p = 0.003). High VEGF expression had a 3-yrRFS of 76% vs 82% for low expression (p = 0.09). HIF-1α expression was not associated with RFS. Ki-67 ≥ 3% was associated with worse 3-yrRFS (≥ 3%: 51% vs < 3%: 84%; p < 0.001), as was the presence of increased microvascular density (CD31 > median: 75% vs CD31 < median: 86%; p = 0.04). High STAT3 expressing tumors were more likely to have a Ki-67 ≥ 3% (42% vs 7%; p < 0.001). LVI was present in 82% of high STAT3 tumors compared to only 50% with low STAT3 (p = 0.058). CD31 expression was similar between groups (58% vs 49%; p = 0.5).

CONCLUSIONS

In resected GEP-NETs, high STAT3 expression is associated with an increased Ki-67 index, presence of lymphovascular invasion and worse 3-yr RFS. STAT3 may be a novel therapeutic target for patients undergoing resection of high-risk tumors.

Association of DCBLD2 upregulation with tumor progression and poor survival in colorectal cancer

PURPOSE

DCBLD2 expression dysregulation has been reported in several types of human cancer. As yet, however, the role of DCBLD2 in colorectal cancer (CRC) is not known.

METHODS

CRC tissues were obtained from patients undergoing surgery from February 2009 to May 2014 (n = 90). Tissue microarray construction and immunohistochemistry were carried out to determine DCBLD2 expression. In vivo studies were performed in 4-week-old BALB/c nude mice. In vitro studies were conducted using CRC-derived HT29 and HCT116 cell lines.

RESULTS

DCBLD2 expression was found to be significantly increased in CRC tissues compared to adjacent normal tissues (p < 0.001). In addition, we found that DCBLD2 expression was positively correlated with the stage of the disease, the degree of differentiation and vascular invasion. High DCBLD2 expression was significantly associated with a poor overall survival. In vitro, DCBLD2 expression downregulation significantly reduced CRC cell proliferation and invasion. In a mouse xenograft model, DCBLD2 expression downregulation reduced lung metastasis and increased overall survival. Gene set enrichment analysis (GSEA) revealed that DCBLD2 overexpression induces epithelial-mesenchymal transition (EMT) and activates the JAK/STAT3 pathway.

CONCLUSIONS

We found that high DCBLD2 expression correlated with a poor clinical outcome, as well as tumorigenesis, invasion and metastasis of CRC cells. DCBLD2 may serve as a prognostic biomarker and a novel therapeutic target for CRC.

Proteomic research in sarcomas - current status and future opportunities

Sarcomas are a rare group of mesenchymal cancers comprising over 70 different histological subtypes. For the majority of these diseases, the molecular understanding of the basis of their initiation and progression remains unclear. As such, limited clinical progress in prognosis or therapeutic regimens have been made over the past few decades. Proteomics techniques are being increasingly utilised in the field of sarcoma research. Proteomic research efforts have thus far focused on histological subtype characterisation for the improvement of biological understanding, as well as for the identification of candidate diagnostic, predictive, and prognostic biomarkers for use in clinic. However, the field itself is in its infancy, and none of these proteomic research findings have been translated into the clinic. In this review, we provide a brief overview of the proteomic strategies that have been employed in sarcoma research. We evaluate key proteomic studies concerning several rare and ultra-rare sarcoma subtypes including, gastrointestinal stromal tumours, osteosarcoma, liposarcoma, leiomyosarcoma, malignant rhabdoid tumours, Ewing sarcoma, myxofibrosarcoma, and alveolar soft part sarcoma. Consequently, we illustrate how routine implementation of proteomics within sarcoma research, integration of proteomics with other molecular profiling data, and incorporation of proteomics into clinical trial studies has the potential to propel the biological and clinical understanding of this group of complex rare cancers moving forward.

Antagonism between the RNA-binding protein Musashi1 and miR-137 and its potential impact on neurogenesis and glioblastoma development

RNA-binding proteins (RBPs) and miRNAs are critical gene expression regulators that interact with one another in cooperative and antagonistic fashions. We identified Musashi1 (Msi1) and miR-137 as regulators of a molecular switch between self-renewal and differentiation. Msi1 and miR-137 have opposite expression patterns and functions, and Msi1 is repressed by miR-137. Msi1 is a stem-cell protein implicated in self-renewal while miR-137 functions as a proneuronal differentiation miRNA. In gliomas, miR-137 functions as a tumor suppressor while Msi1 is a prooncogenic factor. We suggest that the balance between Msi1 and miR-137 is a key determinant in cell fate decisions and disruption of this balance could contribute to neurodegenerative diseases and glioma development. Genomic analyses revealed that Msi1 and miR-137 share 141 target genes associated with differentiation, development, and morphogenesis. Initial results pointed out that these two regulators have an opposite impact on the expression of their target genes. Therefore, we propose an antagonistic model in which this network of shared targets could be either repressed by miR-137 or activated by Msi1, leading to different outcomes (self-renewal, proliferation, tumorigenesis).

The DCBLD receptor family: emerging signaling roles in development, homeostasis and disease

The discoidin, CUB, and LCCL domain-containing (DCBLD) receptor family are composed of the type-I transmembrane proteins DCBLD1 and DCBLD2 (also ESDN and CLCP1). These proteins are highly conserved across vertebrates and possess similar domain structure to that of neuropilins, which act as critical co-receptors in developmental processes. Although DCBLD1 remains largely uncharacterized, the functional and mechanistic roles of DCBLD2 are emerging. This review provides a comprehensive discussion of this presumed receptor family, ranging from structural and signaling aspects to their associations with cancer, physiology, and development.

Distinct genome-wide methylation patterns in sporadic and hereditary nonfunctioning pancreatic neuroendocrine tumors

BACKGROUND

Aberrant methylation is a known cause of cancer initiation and/or progression. There are scant data on the genome-wide methylation pattern of nonfunctioning pancreatic neuroendocrine tumors (NFPanNETs) and sporadic and hereditary NFPanNETs.

METHODS

Thirty-three tissue samples were analyzed: they included samples from sporadic (n = 9), von Hippel-Lindau (VHL)-related (n = 10), and multiple endocrine neoplasia type 1 (MEN1)-related NFPanNETs (n = 10) as well as normal islet cells (n = 4) for comparison. Genome-wide CpG methylation profiling was performed with the Infinium MethylationEPIC BeadChip assay and was analyzed with R-based tools.

RESULTS

In unsupervised hierarchical clustering, sporadic and MEN1-related NFPanNETs clustered together, and the VHL group was in a separate cluster. MEN1-related NFPanNETs had a higher rate of hypermethylated CpG sites in comparison with sporadic and VHL-related tumor groups. Differentially methylated region analysis confirmed the higher rate of hypermethylation in MEN1-related tumors. Moreover, in an integrated analysis of gene expression data for the same tumor samples, downregulated gene expression was found in most genes that were hypermethylated. In a CpG island methylator phenotype analysis, 3 genes were identified and confirmed to have downregulated gene expression: secreted frizzle-related protein 5 (SFRP5) in sporadic NFPanNETs and cell division cycle-associated 7-like (CDCA7L) and RNA binding motif 47 (RBM47) in MEN1-related NFPanNETs.

CONCLUSIONS

MEN1 NFPanNETs have a higher rate of geno me-wide hypermethylation than other NFPanNET subtypes. The similarity between the pathways enriched in a methylation analysis of known genes involved in NFPanNET tumorigenesis suggests a key role for aberrant methylation in the pathogenesis of NFPanNETs.

Circular RNA expression profiles during the differentiation of mouse neural stem cells

Background

Circular RNAs (circRNAs) have recently been found to be expressed in human brain tissue, and many lines ofevidence indicate that circRNAs play regulatory roles in neurodevelopment. Proliferation and differentiation of neural stem cells (NSCs) are critical parts during development of central nervous system (CNS).To date, there have been no reports ofcircRNA expression profiles during the differentiation of mouse NSCs. We hypothesizethat circRNAs mayregulate gene expression in the proliferation anddifferentiation of NSCs.

Results

In this study, we obtained NSCs from the wild-type C57BL/6 J mouse fetal cerebral cortex. We extracted total RNA from NSCs in different differentiation stagesand then performed RNA-seq. By analyzing the RNA-Seq data, we found 37circRNAs and 4182 mRNAs differentially expressedduringthe NSC differentiation. Gene Ontology (GO) enrichment analysis of thecognate linear genes of these circRNAsrevealed that some enriched GO terms were related to neural activity. Furthermore, we performed a co-expression network analysis of these differentially expressed circRNAs and mRNAs. The result suggested a stronger GO enrichmentin neural features for both the cognate linear genes of circRNAs and differentially expressed mRNAs.

Conclusion

We performed the first circRNA investigation during the differentiation of mouse NSCs. Wefound that12 circRNAs might have regulatory roles duringthe NSC differentiation, indicating that circRNAs might be modulated during NSC differentiation.Our network analysis suggested the possible complex circRNA-mRNA mechanisms during differentiation, and future experimental workis need to validate these possible mechanisms.

Electronic supplementary material

The online version of this article (10.1186/s12918-018-0651-1) contains supplementary material, which is available to authorized users.

Small cell lung cancer tumors and preclinical models display heterogeneity of neuroendocrine phenotypes

Background

Small cell lung cancer (SCLC) is a deadly, high grade neuroendocrine (NE) tumor without recognized morphologic heterogeneity. However, over 30 years ago we described a SCLC subtype with "variant" morphology which did not express some NE markers and exhibited more aggressive growth.

Methods

To quantitate NE properties of SCLCs, we developed a 50-gene expression-based NE score that could be applied to human SCLC tumors and cell lines, and genetically engineered mouse (GEM) models. We identified high and low NE subtypes of SCLC in all of our sample types, and characterized their properties.

Results

We found that 16% of human SCLC tumors and 10% of SCLC cell lines were of the low NE subtype, as well as cell lines from the GEM model. High NE SCLC lines grew as non-adherent floating aggregates or spheroids while Low NE lines had morphologic features of the variant subtype and grew as loosely attached cells. While the high NE subtype expressed one of the NE lineage master transcription factors ASCL1 or NEUROD1, together with NKX2-1, the entire range of NE markers, and lacked expression of the neuronal and NE repressor REST, the low NE subtype had lost expression of most NE markers, ASCL1, NEUROD1 and NKX2-1 and expressed REST. The low NE subtype had undergone epithelial mesenchymal transition (EMT) and had activated the Notch, Hippo and TGFβ pathways and MYC oncogene . Importantly, the high and low NE group of SCLC lines had similar gene expression profiles as their SCLC tumor counterparts.

Conclusions

SCLC tumors and cell lines can exhibit distinct inter-tumor heterogeneity with respect to expression of NE features. Loss of NE expression results in major alterations in morphology, growth characteristics, and molecular properties. These findings have major clinical implications as the two subtypes are predicted to have very different responses to targeted therapies.

Discoidin, CUB and LCCL domain-containing protein 2 (DCBLD2) is a novel biomarker of myxofibrosarcoma invasion identified by global protein expression profiling

Myxofibrosarcoma (MFS) is a mesenchymal malignancy characterized by frequent recurrence even after radical wide resection. To optimize therapy for MFS patients, we aimed to identify candidate tissue biomarkers of MFS invasion potential. Invasion characteristics of MFS were evaluated by magnetic resonance imaging and protein expression profiling of primary tumor tissues performed using two-dimensional difference gel electrophoresis (2D-DIGE). Protein expression profiles were compared between invasive and non-invasive tumors surgically resected from 11 patients. Among the 3453 protein spots observed, 59 demonstrated statistically significant difference in intensity (≥2-fold) between invasive and non-invasive tumors (p<0.01 by Wilkoxon test), and were identified by mass spectrometry as 47 individual proteins. Among them, we further focused on discoidin, CUB and LCCL domain-containing protein 2 (DCBLD2), a receptor tyrosine kinase with aberrant expression in malignant tumors. Immunohistochemistry analysis of 21 additional MFS cases revealed that higher DCBLD2 expression was significantly associated with invasive properties of tumor cells. DCBLD2 sensitivity and specificity, and positive and negative predictive values for MFS invasion were 69.2%, 87.5%, 90%, and 63.6%, respectively. The expression level of DCBLD2 was consistent in different portions of tumor tissues. Thus, DCBLD2 expression can be a useful biomarker to evaluate invasive properties of MFS. Further validation studies based on multi-institutional collaboration and comprehensive analysis of DCBLD2 biological functions in MFS are required to confirm its prognostic utility for clinical application.

Brain-Expressed X-linked (BEX) proteins in human cancers

The Brain-Expressed X-linked (BEX) family proteins are comprised of five human proteins including BEX1, BEX2, BEX3, BEX4 and BEX5. BEX family proteins are expressed in a wide range of tissues and are known to play a role in neuronal development. Recent studies suggest a role of BEX family proteins in cancers. BEX1 expression is lost in a subgroup of patients with acute myeloid leukemia (AML) and chronic myeloid leukemia (CML). Expression of BEX1 controls cell surface receptor signaling and restores imatinib response in resistant cells. BEX2 is overexpressed in a group of breast cancer patients and also in gliomas. Increased BEX2 expression led to enhanced NF-κB signaling as well as cell proliferation. Although BEX2 acts as tumor promoter in a subset of breast cancer, BEX3 expression displayed an opposite role. Overexpression of BEX3 resulted in inhibition of tumor formation in breast cancer mouse xenograft models. The role of BEX4 and BEX5 in cancer has not yet been defined. Collectively this suggests that BEX family members have distinct roles in cancers. While BEX1 and BEX3 act as tumor suppressors, BEX2 seems to act as an oncogene.

The role of the small GTPase Rab31 in cancer

Members of the small GTPase family Rab are emerging as potentially important factors in cancer development and progression. A good number of Rabs have been implicated or associated with various human cancers, and much recent excitement has been associated with the roles of the Rab11 subfamily member Rab25 and its effector, the Rab coupling protein (RCP), in tumourigenesis and metastasis. In this review, we focus on a Rab5 subfamily member, Rab31, and its implicated role in cancer. Well recognized as a breast cancer marker with good prognostic value, recent findings have provided some insights as to the mechanism underlying Rab31's influence on oncogenesis. Levels of Oestrogen Receptor α (ERα)- responsive Rab31 could be elevated through stabilization of its transcript by the RNA binding protein HuR, or though activation by the oncoprotein mucin1-C (MUC1-C), which forms a transcriptional complex with ERα. Elevated Rab31 stabilizes MUC1-C levels in an auto-inductive loop that could lead to aberrant signalling and gene expression associated with cancer progression. Rab31 and its guanine nucleotide exchange factor GAPex-5 have, however, also been shown to enhance early endosome-late endosome transport and degradation of the epidermal growth factor receptor (EGFR). The multifaceted action and influences of Rab31 in cancer is discussed in the light of its new interacting partners and pathways.

Gene targeting study reveals unexpected expression of brain-expressed X-linked 2 in endocrine and tissue stem/progenitor cells in mice

Identification of genes specifically expressed in stem/progenitor cells is an important issue in developmental and stem cell biology. Genome-wide gene expression analyses in liver cells performed in this study have revealed a strong expression of X-linked genes that include members of the brain-expressed X-linked (Bex) gene family in stem/progenitor cells. Bex family genes are expressed abundantly in the neural cells and have been suggested to play important roles in the development of nervous tissues. However, the physiological role of its individual members and the precise expression pattern outside the nervous system remain largely unknown. Here, we focused on Bex2 and examined its role and expression pattern by generating knock-in mice; the enhanced green fluorescence protein (EGFP) was inserted into the Bex2 locus. Bex2-deficient mice were viable and fertile under laboratory growth conditions showing no obvious phenotypic abnormalities. Through an immunohistochemical analysis and flow cytometry-based approach, we observed unique EGFP reporter expression patterns in endocrine and stem/progenitor cells of the liver, pyloric stomach, and hematopoietic system. Although Bex2 seems to play redundant roles in vivo, these results suggest the significance and potential applications of Bex2 in studies of endocrine and stem/progenitor cells.

EGFR phosphorylation of DCBLD2 recruits TRAF6 and stimulates AKT-promoted tumorigenesis

Aberrant activation of EGFR in human cancers promotes tumorigenesis through stimulation of AKT signaling. Here, we determined that the discoidina neuropilin-like membrane protein DCBLD2 is upregulated in clinical specimens of glioblastomas and head and neck cancers (HNCs) and is required for EGFR-stimulated tumorigenesis. In multiple cancer cell lines, EGFR activated phosphorylation of tyrosine 750 (Y750) of DCBLD2, which is located within a recently identified binding motif for TNF receptor-associated factor 6 (TRAF6). Consequently, phosphorylation of DCBLD2 Y750 recruited TRAF6, leading to increased TRAF6 E3 ubiquitin ligase activity and subsequent activation of AKT, thereby enhancing EGFR-driven tumorigenesis. Moreover, evaluation of patient samples of gliomas and HNCs revealed an association among EGFR activation, DCBLD2 phosphorylation, and poor prognoses. Together, our findings uncover a pathway in which DCBLD2 functions as a signal relay for oncogenic EGFR signaling to promote tumorigenesis and suggest DCBLD2 and TRAF6 as potential therapeutic targets for human cancers that are associated with EGFR activation.

Gene network-based analysis identifies two potential subtypes of small intestinal neuroendocrine tumors

BACKGROUND

Tumor transcriptomes contain information of critical value to understanding the different capacities of a cell at both a physiological and pathological level. In terms of clinical relevance, they provide information regarding the cellular "toolbox" e.g., pathways associated with malignancy and metastasis or drug dependency. Exploration of this resource can therefore be leveraged as a translational tool to better manage and assess neoplastic behavior. The availability of public genome-wide expression datasets, provide an opportunity to reassess neuroendocrine tumors at a more fundamental level. We hypothesized that stringent analysis of expression profiles as well as regulatory networks of the neoplastic cell would provide novel information that facilitates further delineation of the genomic basis of small intestinal neuroendocrine tumors.

RESULTS

We re-analyzed two publically available small intestinal tumor transcriptomes using stringent quality control parameters and network-based approaches and validated expression of core secretory regulatory elements e.g., CPE, PCSK1, secretogranins, including genes involved in depolarization e.g., SCN3A, as well as transcription factors associated with neurodevelopment (NKX2-2, NeuroD1, INSM1) and glucose homeostasis (APLP1). The candidate metastasis-associated transcription factor, ST18, was highly expressed (>14-fold, p < 0.004). Genes previously associated with neoplasia, CEBPA and SDHD, were decreased in expression (-1.5 - -2, p < 0.02). Genomic interrogation indicated that intestinal tumors may consist of two different subtypes, serotonin-producing neoplasms and serotonin/substance P/tachykinin lesions. QPCR validation in an independent dataset (n = 13 neuroendocrine tumors), confirmed up-regulated expression of 87% of genes (13/15).

CONCLUSIONS

An integrated cellular transcriptomic analysis of small intestinal neuroendocrine tumors identified that they are regulated at a developmental level, have key activation of hypoxic pathways (a known regulator of malignant stem cell phenotypes) as well as activation of genes involved in apoptosis and proliferation. Further refinement of these analyses by RNAseq studies of large-scale databases will enable definition of individual master regulators and facilitate the development of novel tissue and blood-based tools to better understand diagnose and treat tumors.

C18 ORF1, a novel negative regulator of transforming growth factor-β signaling

Transforming growth factor (TGF)-β signaling is deliberately regulated at multiple steps in its pathway from the extracellular microenvironment to the nucleus. However, how TGF-β signaling is activated or attenuated is not fully understood. We recently identified transmembrane prostate androgen-induced RNA (TMEPAI), which is involved in a negative feedback loop of TGF-β signaling. When we searched for a family molecule(s) for TMEPAI, we found C18ORF1, which, like TMEPAI, possesses two PY motifs and one Smad-interacting motif (SIM) domain. As expected, C18ORF1 could block TGF-β signaling but not bone morphogenetic protein signaling. C18ORF1 bound to Smad2/3 via its SIM and competed with the Smad anchor for receptor activation for Smad2/3 binding to attenuate recruitment of Smad2/3 to the TGF-β type I receptor (also termed activin receptor-like kinase 5 (ALK5)), in a similar fashion to TMEPAI. Knockdown of C18ORF1 prolonged duration of TGF-β-induced Smad2 phosphorylation and concomitantly potentiated the expression of JunB, p21, and TMEPAI mRNAs induced by TGF-β. Consistently, TGF-β-induced cell migration was enhanced by the knockdown of C18ORF1. These results indicate that the inhibitory function of C18ORF1 on TGF-β signaling is similar to that of TMEPAI. However, in contrast to TMEPAI, C18ORF1 was not induced upon TGF-β signaling. Thus, we defined C18ORF1 as a surveillant of steady state TGF-β signaling, whereas TMEPAI might help C18ORF1 to inhibit TGF-β signaling in a coordinated manner when cells are stimulated with high levels of TGF-β.

Gene expression differences predict treatment outcome of merkel cell carcinoma patients

Due to the rarity of Merkel cell carcinoma (MCC), prospective clinical trials have not been practical. This study aimed to identify biomarkers with prognostic significance. While sixty-two patients were identified who were treated for MCC at our institution, only seventeen patients had adequate formalin-fixed paraffin-embedded archival tissue and followup to be included in the study. Patients were stratified into good, moderate, or poor prognosis. Laser capture microdissection was used to isolate tumor cells for subsequent RNA isolation and gene expression analysis with Affymetrix GeneChip Human Exon 1.0 ST arrays. Among the 191 genes demonstrating significant differential expression between prognostic groups, keratin 20 and neurofilament protein have previously been identified in studies of MCC and were significantly upregulated in tumors from patients with a poor prognosis. Immunohistochemistry further established that keratin 20 was overexpressed in the poor prognosis tumors. In addition, novel genes of interest such as phospholipase A2 group X, kinesin family member 3A, tumor protein D52, mucin 1, and KIT were upregulated in specimens from patients with poor prognosis. Our pilot study identified several gene expression differences which could be used in the future as prognostic biomarkers in MCC patients.

Systems biology approach to stage-wise characterization of epigenetic genes in lung adenocarcinoma

Background

Epigenetics refers to the reversible functional modifications of the genome that do not correlate to changes in the DNA sequence. The aim of this study is to understand DNA methylation patterns across different stages of lung adenocarcinoma (LUAD).

Results

Our study identified 72, 93 and 170 significant DNA methylated genes in Stages I, II and III respectively. A set of common 34 significant DNA methylated genes located in the promoter section of the true CpG islands were found across stages, and these were: HOX genes, FOXG1, GRIK3, HAND2, PRKCB, etc. Of the total significant DNA methylated genes, 65 correlated with transcription function. The epigenetic analysis identified the following novel genes across all stages: PTGDR, TLX3, and POU4F2. The stage-wise analysis observed the appearance of NEUROG1 gene in Stage I and its re-appearance in Stage III. The analysis showed similar epigenetic pattern across Stage I and Stage III. Pathway analysis revealed important signaling and metabolic pathways of LUAD to correlate with epigenetics. Epigenetic subnetwork analysis identified a set of seven conserved genes across all stages: UBC, KRAS, PIK3CA, PIK3R3, RAF1, BRAF, and RAP1A. A detailed literature analysis elucidated epigenetic genes like FOXG1, HLA-G, and NKX6-2 to be known as prognostic targets.

Conclusion

Integrating epigenetic information for genes with expression data can be useful for comprehending in-depth disease mechanism and for the ultimate goal of better target identification.

Genotype and tumor locus determine expression profile of pseudohypoxic pheochromocytomas and paragangliomas

Pheochromocytomas (PHEOs) and paragangliomas (PGLs) related to mutations in the mitochondrial succinate dehydrogenase (SDH) subunits A, B, C, and D, SDH complex assembly factor 2, and the von Hippel-Lindau (VHL) genes share a pseudohypoxic expression profile. However, genotype-specific differences in expression have been emerging. Development of effective new therapies for distinctive manifestations, e.g., a high rate of malignancy in SDHB- or predisposition to multifocal PGLs in SDHD patients, mandates improved stratification. To identify mutation/location-related characteristics among pseudohypoxic PHEOs/PGLs, we used comprehensive microarray profiling (SDHB: n = 18, SDHD-abdominal/thoracic (AT): n = 6, SDHD-head/neck (HN): n = 8, VHL: n = 13). To avoid location-specific bias, typical adrenal medulla genes were derived from matched normal medullas and cortices (n = 8) for data normalization. Unsupervised analysis identified two dominant clusters, separating SDHB and SDHD-AT PHEOs/PGLs (cluster A) from VHL PHEOs and SDHD-HN PGLs (cluster B). Supervised analysis yielded 6937 highly predictive genes (misclassification error rate of 0.175). Enrichment analysis revealed that energy metabolism and inflammation/fibrosis-related genes were most pronouncedly changed in clusters A and B, respectively. A minimum subset of 40 classifiers was validated by quantitative real-time polymerase chain reaction (quantitative real-time polymerase chain reaction vs. microarray: r = 0.87). Expression of several individual classifiers was identified as characteristic for VHL and SDHD-HN PHEOs and PGLs. In the present study, we show for the first time that SDHD-HN PGLs share more features with VHL PHEOs than with SDHD-AT PGLs. The presented data suggest novel subclassification of pseudohypoxic PHEOs/PGLs and implies cluster-specific pathogenic mechanisms and treatment strategies.

Two new cases of anti-Ca (anti-ARHGAP26/GRAF) autoantibody-associated cerebellar ataxia

Recently, we discovered a novel serum and cerebrospinal fluid (CSF) autoantibody (anti-Ca) to Purkinje cells in a patient with autoimmune cerebellar ataxia (ACA) and identified the RhoGTPase-activating protein 26 (ARHGAP26; alternative designations include GTPase regulator associated with focal adhesion kinase pp125, GRAF, and oligophrenin-1-like protein, OPHN1L) as the target antigen. Here, we report on two new cases of ARHGAP26 autoantibody-positive ACA that were first diagnosed after publication of the index case study. While the index patient developed ACA following an episode of respiratory infection with still no evidence for malignancy 52 months after onset, neurological symptoms heralded ovarian cancer in one of the patients described here. Our finding of anti-Ca/anti-ARHGAP26 antibodies in two additional patients supports a role of autoimmunity against ARHGAP26 in the pathogenesis of ACA. Moreover, the finding of ovarian cancer in one of our patients suggests that anti-Ca/anti-ARHGAP26-positive ACA might be of paraneoplastic aetiology in some cases. In conclusion, testing for anti-Ca/anti-ARHGAP26 should be included in the diagnostic work-up of patients with ACA, and an underlying tumour should be considered in patients presenting with anti-Ca/ARHGAP26 antibody-positive ACA.

Bex2 regulates cell proliferation and apoptosis in malignant glioma cells via the c-Jun NH2-terminal kinase pathway

The function of Bex2, a member of the Brain Expressed X-linked gene family, in glioma is controversial and its mechanism is largely unknown. We report here that Bex2 regulates cell proliferation and apoptosis in malignant glioma cells via the c-Jun NH2-terminal kinase (JNK) pathway. The expression level of Bex2 is markedly increased in glioma tissues. We observed that Bex2 over-expression promotes cell proliferation, while down-regulation of Bex2 inhibits cell growth. Furthermore, Bex2 down-regulation promotes cell apoptosis and activates the JNK pathway; these effects were abolished by administration of the JNK specific inhibitor, SP600125. Thus, Bex2 may be an important player during the development of glioma.

Identification of differentially expressed genes in SHSY5Y cells exposed to okadaic acid by suppression subtractive hybridization

Background

Okadaic acid (OA), a toxin produced by several dinoflagellate species is responsible for frequent food poisonings associated to shellfish consumption. Although several studies have documented the OA effects on different processes such as cell transformation, apoptosis, DNA repair or embryogenesis, the molecular mechanistic basis for these and other effects is not completely understood and the number of controversial data on OA is increasing in the literature.

Results

In this study, we used suppression subtractive hybridization in SHSY5Y cells to identify genes that are differentially expressed after OA exposure for different times (3, 24 and 48 h). A total of 247 subtracted clones which shared high homology with known genes were isolated. Among these, 5 specific genes associated with cytoskeleton and neurotransmission processes (NEFM, TUBB, SEPT7, SYT4 and NPY) were selected to confirm their expression levels by real-time PCR. Significant down-regulation of these genes was obtained at the short term (3 and 24 h OA exposure), excepting for NEFM, but their expression was similar to the controls at 48 h.

Conclusions

From all the obtained genes, 114 genes were up-regulated and 133 were down-regulated. Based on the NCBI GenBank and Gene Ontology databases, most of these genes are involved in relevant cell functions such as metabolism, transport, translation, signal transduction and cell cycle. After quantitative PCR analysis, the observed underexpression of the selected genes could underlie the previously reported OA-induced cytoskeleton disruption, neurotransmission alterations and in vivo neurotoxic effects. The basal expression levels obtained at 48 h suggested that surviving cells were able to recover from OA-caused gene expression alterations.

BEX2 has a functional interplay with c-Jun/JNK and p65/RelA in breast cancer

Background

We have previously demonstrated that BEX2 is differentially expressed in breast tumors and has a significant role in promoting cell survival and growth in breast cancer cells. BEX2 expression protects breast cancer cells against mitochondrial apoptosis and G1 cell cycle arrest. In this study we investigated the transcriptional regulation of BEX2 and feedback mechanisms mediating the cellular function of this gene in breast cancer.

Results

We found a marked induction of BEX2 promoter by c-Jun and p65/RelA using luciferase reporter assays in MCF-7 cells. Furthermore, we confirmed the binding of c-Jun and p65/RelA to the BEX2 promoter using a chromatin immunoprecipitation assay. Importantly, transfections of c-Jun or p65/RelA in MCF-7 cells markedly increased the expression of BEX2 protein. Overall, these results demonstrate that BEX2 is a target gene for c-Jun and p65/RelA in breast cancer. These findings were further supported by the presence of a strong correlation between BEX2 and c-Jun expression levels in primary breast tumors. Next we demonstrated that BEX2 has a feedback mechanism with c-Jun and p65/RelA in breast cancer. In this process BEX2 expression is required for the normal phosphorylation of p65 and IκBα, and the activation of p65. Moreover, it is necessary for the phosphorylation of c-Jun and JNK kinase activity in breast cancer cells. Furthermore, using c-Jun stable lines we showed that BEX2 expression is required for c-Jun mediated induction of cyclin D1 and cell proliferation. Importantly, BEX2 down-regulation resulted in a significant increase in PP2A activity in c-Jun stable lines providing a possible underlying mechanism for the regulatory effects of BEX2 on c-Jun and JNK.

Conclusions

This study shows that BEX2 has a functional interplay with c-Jun and p65/RelA in breast cancer. In this process BEX2 is a target gene for c-Jun and p65/RelA and in turn regulates the phosphorylation/activity of these proteins. These suggest that BEX2 is involved in a novel feedback mechanism with significant implications for the biology of breast cancer.