Robustness of differential gene expression analysis of RNA-seq.

RNA-sequencing (RNA-seq) is a relatively new technology that lacks standardisation. RNA-seq can be used for Differential Gene Expression (DGE) analysis, however, no consensus exists as to which methodology ensures robust and reproducible results. Indeed, it is broadly acknowledged that DGE methods provide disparate results. Despite obstacles, RNA-seq assays are in advanced development for clinical use but further optimisation will be needed. Herein, five DGE models (DESeq2, voom + limma, edgeR, EBSeq, NOISeq) for gene-level detection were investigated for robustness to sequencing alterations using a controlled analysis of fixed count matrices. Two breast cancer datasets were analysed with full and reduced sample sizes. DGE model robustness was compared between filtering regimes and for different expression levels (high, low) using unbiased metrics. Test sensitivity estimated as relative False Discovery Rate (FDR), concordance between model outputs and comparisons of a 'population' of slopes of relative FDRs across different library sizes, generated using linear regressions, were examined. Patterns of relative DGE model robustness proved dataset-agnostic and reliable for drawing conclusions when sample sizes were sufficiently large. Overall, the non-parametric method NOISeq was the most robust followed by edgeR, voom, EBSeq and DESeq2. Our rigorous appraisal provides information for method selection for molecular diagnostics. Metrics may prove useful towards improving the standardisation of RNA-seq for precision medicine.

QuPath: The global impact of an open source digital pathology system.

QuPath, originally created at the Centre for Cancer Research & Cell Biology at Queen's University Belfast as part of a research programme in digital pathology (DP) funded by Invest Northern Ireland and Cancer Research UK, is arguably the most wildly used image analysis software program in the world. On the back of the explosion of DP and a need to comprehensively visualise and analyse whole slides images (WSI), QuPath was developed to address the many needs associated with tissue based image analysis; these were several fold and, predominantly, translational in nature: from the requirement to visualise images containing billions of pixels from files several GBs in size, to the demand for high-throughput reproducible analysis, which the paradigm of routine visual pathological assessment continues to struggle to deliver. Resultantly, large-scale biomarker quantification must increasingly be augmented with DP. Here we highlight the impact of the open source Quantitative Pathology & Bioimage Analysis DP system since its inception, by discussing the scope of scientific research in which QuPath has been cited, as the system of choice for researchers.

Diagnostic accuracy of haptoglobin within ovarian cyst fluid as a potential point-of-care test for epithelial ovarian cancer: an observational study.

OBJECTIVE: To investigate haptoglobin within ovarian cyst fluid (OCF) as a diagnostic biomarker for epithelial ovarian cancer (EOC) and develop an in vitro diagnostic point-of-care device test (IVDPCT) for use in the operating theatre. DESIGN: Retrospective and prospective cohort study. SETTING: South-East Asia. POPULATION: Women with suspicious ovarian cysts. METHODS: Proteomic, immunohistochemical and ELISA methods measured haptoglobin in OCF to differentiate benign and EOCs. Diagnostic performance of haptoglobin was compared with CA125, risk malignancy indices (RMI) and frozen section. Blinded validation of the IVDPCT was performed. MAIN OUTCOME MEASURES: Prediction of malignancy. RESULTS: Haptoglobin concentration measured by ELISA was 0.70 ± 0.09 mg/ml in patients with benign cysts (n = 87), 6.22 ± 0.53 mg/ml in early stage-EOC (n = 17), and 6.57 ± 0.65 mg/ml in late stage-EOC (n = 20). Haptoglobin in EOCs was significantly higher than in benign cysts (P < 0.0001). Haptoglobin using rapid colorimetric assay (RCA) on a training set had a sensitivity of 97.3% and a specificity 92.0%, comparable to ELISA and frozen sections. The haptoglobin AUROC curve was 0.999 (95% CI 0.997-1.000) compared with 0.895 (95% CI 0.814-0.977, P < 0.05) for CA125. Haptoglobin performed significantly better than all the RMIs (P < 0.01). Blinded validation studies showed a minor drop in average diagnostic performance (sensitivity 85.2% and specificity 90.5%) compared with the training set. However, when compared with frozen section, haptoglobin was no worse in diagnostic accuracy for malignancy. CONCLUSION: Haptoglobin was identified as a biomarker for the detection of EOC with potential as a point-of-care diagnostic tool. TWEETABLE ABSTRACT: Haptoglobin within ovarian cyst fluid: a biomarker for epithelial ovarian cancer and point-of-care diagnostics.

Training in molecular cytopathology testing.

Training in molecular cytopathology testing is essential in developing and maintaining skills in modern molecular technologies as they are introduced to a universal health care system such as extant in the UK and elsewhere. We review the system in place in Northern Ireland (NI) for molecular testing of solid tumours, as an example to train staff of all grades, including pathologists, clinical scientists, biomedical scientists and equivalent technical grades. We describe training of pathologists as part of the NI Deanery medical curriculum, the NI training programme for scientists and laboratory rotation for Biomedical Scientists. Collectively, the aims of our training are two-fold: to provide a means by which individuals may extend their experience and skills; and to provide and maintain a skilled workforce for service delivery. Through training and competency, we introduce new technologies and tests in response to personalised medicine therapies with a competent workforce. We advocate modifying programmes to suit individual needs for skill development, with formalised courses in pre-analytical, analytical and postanalytical demands of modern molecular pathology. This is of particular relevance for cytopathology in small samples such those from formalin-fixed paraffin-embedded cell blocks. We finally introduce how university courses can augment training and develop a skilled workforce to benefit the delivery of services to our patients.

RUNX3 is oncogenic in natural killer/T-cell lymphoma and is transcriptionally regulated by MYC.

RUNX3, runt-domain transcription factor, is a master regulator of gene expression in major developmental pathways. It acts as a tumor suppressor in many cancers but is oncogenic in certain tumors. We observed upregulation of RUNX3 mRNA and protein expression in nasal-type extranodal natural killer (NK)/T-cell lymphoma (NKTL) patient samples and NKTL cell lines compared to normal NK cells. RUNX3 silenced NKTL cells showed increased apoptosis and reduced cell proliferation. Potential binding sites for MYC were identified in the RUNX3 enhancer region. Chromatin immunoprecipitation-quantitative PCR revealed binding activity between MYC and RUNX3. Co-transfection of the MYC expression vector with RUNX3 enhancer reporter plasmid resulted in activation of RUNX3 enhancer indicating that MYC positively regulates RUNX3 transcription in NKTL cell lines. Treatment with a small-molecule MYC inhibitor (JQ1) caused significant downregulation of MYC and RUNX3, leading to apoptosis in NKTL cells. The growth inhibition resulting from depletion of MYC by JQ1 was rescued by ectopic MYC expression. In summary, our study identified RUNX3 overexpression in NKTL with functional oncogenic properties. We further delineate that MYC may be an important upstream driver of RUNX3 upregulation and since MYC is upregulated in NKTL, further study on the employment of MYC inhibition as a therapeutic strategy is warranted.

Integrating molecular diagnostics into histopathology training: the Belfast model.

Molecular medicine is transforming modern clinical practice, from diagnostics to therapeutics. Discoveries in research are being incorporated into the clinical setting with increasing rapidity. This transformation is also deeply changing the way we practise pathology. The great advances in cell and molecular biology which have accelerated our understanding of the pathogenesis of solid tumours have been embraced with variable degrees of enthusiasm by diverse medical professional specialties. While histopathologists have not been prompt to adopt molecular diagnostics to date, the need to incorporate molecular pathology into the training of future histopathologists is imperative. Our goal is to create, within an existing 5-year histopathology training curriculum, the structure for formal substantial teaching of molecular diagnostics. This specialist training has two main goals: (1) to equip future practising histopathologists with basic knowledge of molecular diagnostics and (2) to create the option for those interested in a subspecialty experience in tissue molecular diagnostics to pursue this training. It is our belief that this training will help to maintain in future the role of the pathologist at the centre of patient care as the integrator of clinical, morphological and molecular information.

Fibroblast growth factor receptor 4 (FGFR4): a targetable regulator of drug resistance in colorectal cancer.

The discovery of underlying mechanisms of drug resistance, and the development of novel agents to target these pathways, is a priority for patients with advanced colorectal cancer (CRC). We previously undertook a systems biology approach to design a functional genomic screen and identified fibroblast growth factor receptor 4 (FGFR4) as a potential mediator of drug resistance. The aim of this study was to examine the role of FGFR4 in drug resistance using RNAi and the small-molecule inhibitor BGJ398 (Novartis). We found that FGFR4 is highly expressed at the RNA and protein levels in colon cancer tumour tissue compared with normal colonic mucosa and other tumours. Silencing of FGFR4 reduced cell viability in a panel of colon cancer cell lines and increased caspase-dependent apoptosis. A synergistic interaction was also observed between FGFR4 silencing and 5-fluorouracil (5-FU) and oxaliplatin chemotherapy in colon cancer cell lines. Mechanistically, FGFR4 silencing decreased activity of the pro-survival STAT3 transcription factor and expression of the anti-apoptotic protein c-FLIP. Furthermore, silencing of STAT3 resulted in downregulation of c-FLIP protein expression, suggesting that FGFR4 may regulate c-FLIP expression via STAT3. A similar phenotype and downstream pathway changes were observed following FGFR4 silencing in cell lines resistant to 5-FU, oxaliplatin and SN38 and upon exposure of parental cells to the FGFR small-molecule inhibitor BGJ398. Our results indicate that FGFR4 is a targetable regulator of chemo-resistance in CRC, and hence inhibiting FGFR4 in combination with 5-FU and oxaliplatin is a potential therapeutic strategy for this disease.

Prognostic and therapeutic relevance of FLIP and procaspase-8 overexpression in non-small cell lung cancer.

Non-small cell lung carcinoma remains by far the leading cause of cancer-related deaths worldwide. Overexpression of FLIP, which blocks the extrinsic apoptotic pathway by inhibiting caspase-8 activation, has been identified in various cancers. We investigated FLIP and procaspase-8 expression in NSCLC and the effect of HDAC inhibitors on FLIP expression, activation of caspase-8 and drug resistance in NSCLC and normal lung cell line models. Immunohistochemical analysis of cytoplasmic and nuclear FLIP and procaspase-8 protein expression was carried out using a novel digital pathology approach. Both FLIP and procaspase-8 were found to be significantly overexpressed in tumours, and importantly, high cytoplasmic expression of FLIP significantly correlated with shorter overall survival. Treatment with HDAC inhibitors targeting HDAC1-3 downregulated FLIP expression predominantly via post-transcriptional mechanisms, and this resulted in death receptor- and caspase-8-dependent apoptosis in NSCLC cells, but not normal lung cells. In addition, HDAC inhibitors synergized with TRAIL and cisplatin in NSCLC cells in a FLIP- and caspase-8-dependent manner. Thus, FLIP and procaspase-8 are overexpressed in NSCLC, and high cytoplasmic FLIP expression is indicative of poor prognosis. Targeting high FLIP expression using HDAC1-3 selective inhibitors such as entinostat to exploit high procaspase-8 expression in NSCLC has promising therapeutic potential, particularly when used in combination with TRAIL receptor-targeted agents.

Prolyl isomerase Pin1 downregulates tumor suppressor RUNX3 in breast cancer.

Emerging evidence demonstrates that RUNX3 is a tumor suppressor in breast cancer. Inactivation of RUNX3 in mice results in spontaneous mammary gland tumors, and decreased or silenced expression of RUNX3 is frequently found in breast cancer cell lines and human breast cancer samples. However, the underlying mechanism for initiating RUNX3 inactivation in breast cancer remains elusive. Here, we identify prolyl isomerase Pin1, which is often overexpressed in breast cancer, as a key regulator of RUNX3 inactivation. In human breast cancer cell lines and breast cancer samples, expression of Pin1 inversely correlates with the expression of RUNX3. In addition, Pin1 recognizes four phosphorylated Ser/Thr-Pro motifs in RUNX3 via its WW domain. Binding of Pin1 to RUNX3 suppresses the transcriptional activity of RUNX3. Furthermore, Pin1 reduces the cellular levels of RUNX3 in an isomerase activity-dependent manner by inducing the ubiquitination and proteasomal degradation of RUNX3. Knocking down Pin1 enhances the cellular levels and transcriptional activity of RUNX3 by inhibiting the ubiquitination and degradation of RUNX3. Our results identify Pin1 as a new regulator of RUNX3 inactivation in breast cancer.

The positive impact of cytological specimens for EGFR mutation testing in non-small cell lung cancer: a single South East Asian laboratory's analysis of 670 cases.

OBJECTIVES: To compare the rejection rates of non-small cell lung cancer (NSCLC) samples obtained by differing sampling methods for testing by Sanger sequencing for epidermal growth factor receptor (EGFR) mutations. To assess the association between unsatisfactory outcomes and the quantity of DNA extracted from cytological versus histological samples. METHODS: Six hundred and seventy NSCLC samples referred to our centre from 2008 to 2010 were reviewed as a consequence of sample rejection, presence of EGFR mutations, cytological versus histological sampling methods, DNA quantity and the unsatisfactory genotyping rate. RESULTS: Eighty samples were rejected for testing in similar proportions of histological and cytological samples (11.9% versus 10.9%) usually (n = 75) because the amount of cellular material was judged insufficient in small biopsies or cytology samples. The remaining 590 samples on which EGFR testing was attempted yielded 51 (8.6%) unsatisfactory test outcomes caused by failure of the polymerase chain reaction (PCR) (n = 47 cases), uninterpretable Sanger chromatograms (n = 3 cases) and insufficient DNA extracted for PCR (n = 1 case). The difference in rates of unsatisfactory outcomes between cytological samples (seven of 147 samples or 4.7%) versus tissue samples (44 of 443 samples or 9.9%) was clinically relevant but not statistically significant (Mann-Whitney test; P < 0.081). There was no association between the concentration of DNA extracted and the likelihood of an unsatisfactory analysis; which was similar in all types of sections (large and small) while 0% of 37 cytology slides were unsatisfactory. CONCLUSIONS: Utilizing cytology samples for EGFR testing avoids unnecessary patient re-biopsing and yields a clinically superior satisfactory rate to the overall satisfactory rate of tissue biopsies of NSCLC. The quality rather than quantity of DNA extracted may be a more important determinant of a satisfactory result.

A morpho-molecular prognostic model for hepatocellular carcinoma.

BACKGROUND: Hepatocellular carcinoma (HCC) is the third common cause of cancer-related deaths and its prognostication is still suboptimal. The aim of this study was to establish a new prognostication algorithm for HCC. METHODS: In all, 13 biomarkers related to the etiopathogenesis of HCC were evaluated by immunohistochemistry using tissue microarrays containing 121 primary HCC resection cases, and validated in subsequent cohort of 85 HCC cases. The results were compared with Affymetrix Gene Chip Human Genome U133Plus microarray data in a separate cohort of 228 HCC patients. RESULTS: On immunohistochemical evaluation and multivariate Cox regression analysis p53, alpha fetaprotein (AFP), CD44 and CD31, tumour size and vascular invasion, were significant predictors for worse survival in HCC patients. A morpho-molecular prognostic model (MMPM) was constructed and it was a significant independent predictor for overall survival (OS) and relapse-free survival (RFS) (P<0.000). The OS and RFS of HCC(low) was higher (104 and 78 months) as compared with HCC(high) (73 and 43 months) (P<0.000 for OS and RFS). Hepatocellular carcinoma patients with higher stage (III+IV), >5 cm tumour size, positive vascular invasion and satellitosis belonged to HCC(high) group. The validation group reproduced the same findings. Gene expression analysis confirmed that 7 of the 12 biomarkers were overexpressed in >50% of tumour samples and significant overexpression in tumour samples was observed in AFP, CD31, CD117 and Ki-67 genes. CONCLUSION: The MMPM, based on the expression of selected proteins and clinicopathological parameters, can be used to classify HCC patients between good vs poor prognosis and high vs low risk of recurrence following hepatic resection.

RUNX3 acts as a tumor suppressor in breast cancer by targeting estrogen receptor α.

Transcription factor RUNX3 is inactivated in a number of malignancies, including breast cancer, and is suggested to function as a tumor suppressor. How RUNX3 functions as a tumor suppressor in breast cancer remains undefined. Here, we show that about 20% of female Runx3(+/-) mice spontaneously developed ductal carcinoma at an average age of 14.5 months. Additionally, RUNX3 inhibits the estrogen-dependent proliferation and transformation potential of ERα-positive MCF-7 breast cancer cells in liquid culture and in soft agar and suppresses the tumorigenicity of MCF-7 cells in severe combined immunodeficiency mice. Furthermore, RUNX3 inhibits ERα-dependent transactivation by reducing the stability of ERα. Consistent with its ability to regulate the levels of ERα, expression of RUNX3 inversely correlates with the expression of ERα in breast cancer cell lines, human breast cancer tissues and Runx3(+/-) mouse mammary tumors. By destabilizing ERα, RUNX3 acts as a novel tumor suppressor in breast cancer.

Sequential expression of putative stem cell markers in gastric carcinogenesis.

BACKGROUND: Gastric carcinogenesis has been well documented in the step-wise histopathological model, known as Correa pathway. Several biomarkers including CD44, Musashi-1 and CD133 have been reported as putative stem cell (PSC) markers. METHODS: We investigated expression of PSC markers CD44, Musashi-1 and CD133 in relation to gastric carcinogenesis and prognosis and chemoresponse. Immunohistochemistry staining was performed in gastric cancer (GC) clinical specimens representing different steps of the Correa pathway. Gastric cancer samples taken before and after neoadjuvant chemotherapy with docetaxel, cisplatin and capecitabine (DCX) were also evaluated for PSC marker expression. RESULTS: We showed that the expression of three PSC markers was significantly elevated in GC relative to normal gastric mucosa (P<0.001 for each marker). Precancerous lesions, including intestinal metaplasia and dysplasia, demonstrated increased expression of CD44 and Musashi-1. CD133 was predominantly expressed along the border between intramucosal carcinoma and connective tissue at later stages. High CD44 and CD133 expression showed prognostic value for worse patient survival (P=0.014 and P=0.019, respectively). A small number of tumours with high expression of CD44 and CD133 showed pathological response to DCX-based neoadjuvant chemotherapy. CONCLUSION: CD44 and Musashi-1 are frequently expressed in both premalignant gastric lesions and invasive GC, whereas CD133 expression is restricted mainly to neoplastic tissues.

Serial changes in the expression of breast cancer-related proteins in response to neoadjuvant chemotherapy.

BACKGROUND: Tumour expression of cyclooxygenase-2 (COX-2), epidermal growth factor receptor (EGFR), erythroblastic leukaemia viral oncogene homologue-2 (ErbB2), Ki-67 and p53 in breast cancer are associated with poorer outcomes. We investigated in vivo changes of these proteins with neoadjuvant chemotherapy. PATIENTS AND METHODS: Four core biopsies were taken from 100 breast cancer patients at baseline, during and upon completion of neoadjuvant chemotherapy. Immunohistochemical expression of these proteins were evaluated and correlated with clinicopathological features, clinical response and progression-free survival (PFS). RESULTS: There was a statistically significant change from positivity to negativity in COX-2 expression with chemotherapy (P = 0.002), predominantly in clinical responders (P = 0.002). COX-2-positive tumours that remained positive had shorter PFS than those that turned negative. Estrogen receptor (ER)+ and COX-2+ tumours at baseline that remained COX-2+ fared worse than those that became COX-2 negative (PFS 27 versus 52 months, P = 0.002). No significant changes in IHC expression were observed for ER, progesterone receptor, ErbB2, EGFR, p53 or Ki67. CONCLUSIONS: Chemotherapy induced change in COX-2 expression from positivity to negativity predominantly among clinical responders and is associated with longer PFS. Interaction between COX-2 and ER was observed, suggesting that some hormone receptor-positive patients may benefit from combining COX-2 inhibition with hormonal therapy.

KRAS and BRAF mutation analysis can be reliably performed on aspirated cytological specimens of metastatic colorectal carcinoma.

BACKGROUND: Sanger sequencing is one of several reliable methods in use to detect KRAS and BRAF mutations to facilitate clinical patient selection for anti-epidermal growth factor receptor (EGFR) monoclonal antibody therapy in unresectable metastatic colorectal adenocarcinoma (CRC). Most analyses are made on pretreatment biopsy or resection specimens. There is a scarcity of published studies on the suitability of cytological samples for KRAS testing in this setting. METHODS: DNA extraction was attempted on 11 search-retrieved paired cases of histological resections or excisions of CRC and their corresponding cytological samples (representing metastases) and tested for KRAS mutations in exon 2 and 3, as well as BRAF exon 15 mutations by Sanger sequencing. Only KRAS wild-type cases were subjected to BRAF analysis because this is the setting with true diagnostic value, as these mutations are mutually exclusive. RESULTS: Of the 11 paired cases analysed, only eight histology cases showed satisfactory DNA quality for sequencing. Thus, only eight of the corresponding cytology cases were analysed. Seven of the eight cases tested showed the same KRAS genotype on both the aspirated cytology specimen of metastatic carcinoma and the primary tumour (histological specimen), from which we derive an overall concordance rate of 87.5%. The single discordant case was likely to be a true difference as it was demonstrated again on repeat testing of both samples. No BRAF mutations were detected on the four KRAS wild-type cases. CONCLUSION: A range of cytological samples are suitable for KRAS and BRAF mutation testing, be it from previously stained preparations or cell blocks. These samples would be highly valuable in cases where cytological samples are the only material available for mutation testing.

Comparative validation of c-kit exon 11 mutation analysis on cytology samples and corresponding surgical resections of gastrointestinal stromal tumours.

OBJECTIVE: Studies have shown that c-kit mutation analysis of gastrointestinal stromal tumours (GISTs) obtained by endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) can be routinely performed. We validated c-kit exon 11 mutational analysis on cell block material obtained from fine needle aspiration cytology (FNAC) for diagnostic purposes and compared it with the same analysis in formalin-fixed paraffin-embedded full sections of the corresponding resection specimens. METHODS: c-kit mutation analysis was done on cell block material obtained from ten cases encountered in our department from 1999 to 2008 on which FNAC was attempted pre-operatively. The findings were compared with analysis on full paraffin section of the corresponding resected tumours in seven cases where patients opted for resection. c-kit exon 11 was examined via bidirectional nucleic acid sequencing. RESULTS: Our results showed 100% concordance for the presence and type of exon 11 mutation in the resected and aspirated tumours in all seven cases. These mutations had diagnostic value when compared with other neoplasms that are part of the cytomorphological differential diagnosis, such as leiomyosarcoma or gastric adenocarcinomas. CONCLUSION: Molecular cytopathology is a powerful tool that can complement morphology and immunohistochemical assessment of cytological material in routine practice for the diagnosis and prognostication of GISTs. We briefly discuss the advantages and limitations of the fine needle method of obtaining tissue for the diagnosis and prognostication of GISTs, and its current therapeutic strategies.