List of Publications

This page lists my publications currently in print; articles in press, submitted and manuscripts-in-preparation are listed under current projects. Due to copyright restrictions, abstracts are not presented (most are hyperlinked); I have included a short summary of each paper. I've used a bigger font to designate papers of special significance to my current projects and plans and/or to clinical practice.

A major step in proving the clinical validity of a cancer test is confirmation on a second series, and in particular a high level of evidence can be achieved if two similar randomized clinical trials can be brought to bear on the same question and give the same answer. My Danish colleagues had previously developed a predictor of which women will benefit from anthracyclines (an aggressive type of breast cancer chemotherapy). I was able to arrange a collaboration between the major clinical trial groups in Denmark and Canada to confirm this finding (developed on Danish trial DBCG89D) on the independent Canadian trial MA.5. This work is forming the basis for ongoing collaborations between our two groups.
This work describes an algorithm developed by my colleagues at Stanford University for the computerized analysis of digitized breast cancer pathology images. We provide a critical validation series that proved this technique is able to identify features associated with subsequent patient survival. A novel finding is the major contribution of features of the breast cancer stroma, as opposed to the cancer cells themselves.
In support of the IND.200 clinical trial, we gathered together large numbers of sarcomas in a series of tissue microarrays, and optimized immunohistochemical biomarker tests for the major enzymes targeted by the drug under study (SB939, a histone deacetylase inhibitor). We were able to show consistent, high level expression of HDAC2 in fusion transcription-factor associated mesenchymal tumors.
I was invited to present to an international working group on technical aspects of Ki67 testing, at a meeting in London which evolved into this white paper on best practices for Ki67 measurement, interpretation and clinical use. Although this work represents an important synthesis of two decades of work, it is clear there are many unanswered questions about scoring systems and interobserver consistency. As a result, I am directing a follow-up project with my international colleagues, to define additional best practices for clinical validity.
This work represents a multiinstitutional collaborative effort. Our British colleagues trained a prognostic classifier based on simple immunohistochemistry for the proliferation marker Ki67 and the antiapoptotic protein BCL2. We provided a large independent validation set of over 3000 breast cancer samples, that provided the critical proof of the value of this survival predictor among estrogen receptor expressing breast cancers.
My first clinical fellow, Hassan Huwait, worked with me to put together this report that expands the spectrum of histological and clinical presentations of Kaposi sarcoma, and virally-induced vascular neoplasm .
Whose research career can be considered complete without a study on blind dwarves from Newfoundland? Next generation sequencing gives us a chance to identify new genetic abnormalities underlying rare syndromes, as demonstrated elegantly in this study describing the genotype and phenotype induced by specific hereditary mutations in GNPTG, associated with mucolipidosis.

An elegant molecular FISH confirmation, by our resident Anna Lee, confirms that low grade fibromyxoid sarcomas can ossify, producing a deceptive appearance on imaging studies and core biopsy.

I played a major part in the formulation of these guidelines for tissue microarray construction and analysis on clinical trials, drawing on my experience at GPEC and my roles on the CALGB Pathology and NCICCTG Correlative Science executive committees.
In this paper we describe a mathematical algorithmic approach to identifying translocations representing transcribed fusion genes from the immense backdrop of next generation sequencing data, in this case from whole transcriptome shotgun sequencing of many kinds of tumor specimens, including sarcomas. In doing so, we also identify a novel RREB1-TFE3 fusion transcription factor oncogene in a sarcoma from a young adult.
Using our 4000-case British Columbia tissue microarray series, we validate FOXA1 as one of those rare good prognostic factors in breast cancer - associated with a low risk subset of estrogen receptor positive tumors.
I was privileged to be named the 2010 Joe Doupe Young Investigator Award winner, by the Canadian Society for Clinical Investigation. This article elaborates on my award speech, describing the path I took from resident, through basic science discovery work, setting up my own lab and testing hypotheses in vitro, building evidence and eventually initiating a clinical trial (NCIC-CTG IND.200) for sarcoma patients based on my research.

In this work we apply the new PAM50 gene expression assay to a large collection of paraffin blocks linked to detailed clinical and outcome data. We demonstrate the capacity of this assay to identify the molecular subtype of breast cancer, and its superiority over the best clinical and immunohistochemical models for predicting the outcome of women in the common scenario of hormone-receptor positive breast cancer treated with endocrine therapy.

This work, funded by the Sarcoma Foundation of America and the Cancer Research Society, describes the world's first model of tenosynovial giant cell tumor suitable for testing of experimental therapeutics, and demonstrates the effectiveness of antibodies against human CSF1 against this disease. Published in an open access journal.

After an extensive effort to carefully annotate all the information about metastatic site and spread in our British Columbia Breast Cancer Outcomes Unit database, we were able to link intrinsic subtype to patterns of metastatic spread, finding for example that luminal cancers spread preferentially to bone, and basal tumors to lung.

This molecular tour-de-force by my student Le Su makes the link connecting the synovial sarcoma oncogene to abnormally low expression of the critical PTEN tumor suppressor gene, an advance in understanding the molecular oncogenesis of this disease which has implications for possible therapies with investigational new drugs.

This open access case report describes an extremely interesting and deceptive case of clear cell sarcoma.

We were the largest single contributor to this open access meta-analysis of the immunohistochemical subtyping of breast cancer, which confirmed our findings on the prognostic significance of the Luminal A, Luminal B, HER2-associated and basal breast cancer subtypes.
This work discusses the basic philosophy of the sarcoma research programs we are undertaking at UBC and at Stanford, in the context of our current knowledge and work that is going on world-wide. The editor of JCO solicited this review from me, following my presentation at Harvard Medical School in spring of 2009, and I wrote it with my close colleague Rob West from Stanford.
Using our large British Columbia tissue microarray series and our established immunohistochemical panel for breast cancer intrinsic subtyping, we show that the non-Luminal A types of breast cancer are associated with increased risk of both local recurrence and regional lymph node relapse, not only post-lumpectomy, but also post-mastectomy.
Having previously identified TLE1 as a specific and sensitive marker of synovial sarcoma by gene expression and tissue microarray analyses, for the next two year, all incoming cases at the BC Cancer Agency and the Cleveland Clinic where synovial sarcoma was considered in the differential diagnosis were prospectively evaluated for TLE1, keratin, EMA and bcl2 immunostains against a molecular translocation test gold standard. The result: 92% positive predictive value and 100% negative predictive value in a real-life diagnostic situation, cementing its role as the best immunostain for this diagnosis, and obviating the need for more expensive molecular testing in many cases. We believe this will be practice changing.
The result of a multi-institution consortium to study the expression profile of malignant peripheral nerve sheath tumors, funded by the US Department of Defense, we were able to link repression of p53 and specific micro-RNA pathways to malignant progression from benign neurofibromas to malignant tumors.
IGF2 is a critical signalling molecule in growth and development, and has been co-opted by many cancers. New drugs targeting this pathway are becoming available. In this study, we use a large series of tissue microarrays from Canada, the USA and Norway to make a comprehensive survey of IGF2 expression in benign and malignant bone and soft tissue tumors. The highest expression was found in solitary fibrous tumors and synovial sarcomas.
In this paper from our Strategic Partnering to Evaluate Cancer Signatures consortium, with work spearheaded by our GPEC group, we are able to define a cutpoint for the proliferation index which allows the best distinction between indolent luminal A and aggressive luminal B breast cancers to be made by a cheap immunohistochemistry test. This work is published in an on-line format, freely accessible world-wide.
In this work funded from our CIHR grant, we use gene expression profiling data to identify biomarkers in myxoid/round cell liposarcoma, a deceptively aggressive cancer for which no truly effective systemic therapy is available. Using tissue microarrays linked to outcome, we implicate the IGF and RET oncogenic pathways as prognostically important in this disease.
Our star resident research trainee Chenghan Lee did an elective at Stanford, bringing ideas and specimens from Vancouver to help validate the group's previous finding that macrophage signatures are prognostic in leiomyosarcoma, and extend this to show that this occurs via CSF1 production and signaling by tumor cells, recruiting the macrophages that contribute to aggressive disease features.
The value of progesterone receptor testing in breast cancer is controversial. After working our way through many of the commercial antibodies, we were able to show that a new rabbit monoclonal, 1E2, works well on our 4000 case tissue microarray series of British Columbia breast cancers. Furthermore, progesterone receptor testing does recognize a small fraction (4%) of cases which could be missed if estrogen receptor testing was used alone, and adds significant additional prognostic information across treatment groups and in multivariate analyses.
This paper demonstrates the culmination of a lot of our work in the Strategic Partnering to Evaluation Cancer Signatures Program. We define and validate a 50 gene classifier panel that accurately subtypes breast cancers into their major intrinsic subtypes. This panel is prognostic and predictive, and can be assayed by qRT-PCR on standard formalin-fixed, paraffin-embedded laboratory specimens.
Our surrogate immunohistochemical panel for breast cancer intrinsic subtyping was applied to this major international breast cancer trial, and reveals that our classification is predictive of response to docetaxel chemotherapy (which benefits women with all major subtypes except Luminal A breast cancers).
Anne Nguyen's MSc thesis culminated in this open-access paper, where we demonstrate that Hsp90 and HDAC inhibitors have synergistic activity against synovial sarcoma in vitro. Morever, this synergism is likely mediated by the Nf-kappa-B pathway, inhibitors of which we also demonstrate (for the first time) to be active against synovial sarcoma.
In this case report, we expand the spectrum of histological appearances for gastrointestinal stromal tumors, proving the diagnosis with DNA sequencing for activating receptor tyrosine kinase mutations.
Dave Voduc and I were asked to write this review article, and we accepted as there have been several other review articles out there on basal breast cancer, some from people with little real experience in how to interpret the diagnostic aspects of this very aggressive form of breast cancer.
This work represents some of the output from our very active collaboration among the Genetic Pathology Evaluation Centre, the Breast Cancer Outcomes Unit of the BC Cancer Agency, and PhenoPath laboratories. We are able to define the frequency of HER2 overexpression in a large population of early (node negative) breast cancers as 10% - somewhat lower than previous estimates - and show the adverse natural history among a historically-important cohort of patients who did not receive adjuvant systemic therapy.
Working with our close collaborators from the Prostate Centre, who have developed antisense oligonucleotide therapies targeting IGFBP2, we show that this protein is highly expressed in the majority of breast cancers, but not in benign tumors of the breast. This finding suggests that their new antisense drug OGX-225 might have activity against a broad spectrum of breast tumors, affecting a growth pathway co-opted by many cancers.
The oncogene YB-1, a transcriptional regulator of both the EGFR and HER2 breast cancer oncogenes, is associated with poor patient survival, regardless of biological subtype, in our 4000-case British Columbia breast cancer tissue microarray series.
In this paper I contributed data and ideas for a multinational collaborative study confirming the association of the epidermal growth factor receptor oncogene with basal-like breast cancers, particularly those in patients with hereditary BRCA1 forms of aggressive breast cancer.
This one is just a case report that the radiologists put my name on and published without asking me first. Please never do that to me!
We assess, on breast cancer tissue microarrays, the cyclin E - p27Kip1 - Skp2 axis, regulating cell cycle control, and find that by immunohistochemistry it is the two oncogenes which are associated with the aggressive basal-like subtype of breast cancer, and with adverse patient outcomes.
Chenghan Lee, a graduate of the UBC MD/PhD program, followed in my footsteps to do a resident research elective in Matt van de Rijn's lab. Among the fruits of his labour there was this study using gene and tissue microarrays to analyze giant cell tumor of bone. You can find a .pdf of the article here.
This review article, aimed at clinicians, describes tissue microarray technology, its advantages and disadvantages, and how it can be used to accelerate translation of new research into clinically-useful tests.
We demonstrate, on a large population-based tissue microarray series, that almost all the additional risk of the clinically important "triple negative" type of breast cancer is conferred by the subset of "core basal" tumors that express positive markers (EGFR and cytokeratin 5) of basal breast cancer. Furthermore, anthracycline-treated patients may also do relatively poorly if they have this phenotype, which can easily be identified in clinical labs.
Macrophage infiltrates in cancer have been linked with aggressive tumor behaviour. Here we identify a macrophage gene expression signature in a subset of leiomyosarcomas, which is not seen in other sarcomas. With tissue microarrays linked to patient outcome, we confirm this finding in larger numbers of patients and show for the first time that such infiltrates portend poor outcomes in soft tissue (but not uterine) leiomyosarcomas.
Working with Nobel Laureate Andrew Fire, we present the first large microarray study of microRNA expression in human sarcomas
Working with my fellow Dave Voduc and PhD student Maggie Cheang, we use our 4000-case breast cancer tissue microarray to investigate the clinical value of GATA-3, a transcription factor associated with the good prognosis, estrogen receptor positive, Luminal A form of breast cancer.
In continuing followup of our groundbreaking study of sarcoma expression profiling, we develop a diagnostic monoclonal antibody that can be more widely available than our previous polyclonal. DOG1 (aka TMEM16A) is an excellent, sensitive and specific marker for GI stromal tumors, which we are already using diagnostically at Stanford and UBC. It has particular value for kit negative cases.
My student Maggie and I wrote this critical review of the application of gene expression technology to the study of breast cancer, for the February 2008 issue of Annual Reviews of Pathology: Mechanisms of Disease. If you are interested to read it, direct downloads can be found at the following site: http://arjournals.annualreviews.org/doi/pdf/10.1146/annurev.pathmechdis.3.121806.151505
Once our GI stromal tumor tissue array was built, Sonja and I quickly engaged in other studies, including using our BLISS automated image analysis platform to perform a comprehensive study into the value of nuclear morphometry in GI stromal tumors against a gold standard of patient natural history
The results of FISH analysis for HER2 on our large breast cancer tissue microarray series took a great deal of time and effort to obtain, but is of value for several studies. Here, we perform comprehensive cutpoint analysis and find that the current requirement for an amplification ratio of >2.2 for a case to be considered positive for HER2 is probably too stringent
Sonja, my sabbatical fellow from the University of Northern Norway, spent a year with me during which time we built, stained and scored a series of GI stromal tumors representing almost the full population of Norway in the two decades preceding the Gleevec era. In our first publication using this material, we describe the series and how established biomarkers map to patient outcome
Working with my radiology colleagues in the provincial musculoskeletal tumor group, we identify an extreme case of metallosis resulting from breakdown of a failed joint prosthesis, and present the imaging and pathologic findings
Marc van de Vijver's group published a study claiming that gene expression analysis revealed ER/PR/HER2 triple negative breast cancers to have Basal expression profiles. Along with other Canadian and British researchers, we had serious concerns with the supporting date presented and the extent of the conclusions that were drawn. It is clear from our own and others' research that there are in fact many "triple negative" cases that behave better than basal tumors, and that there are basal tumors which are positive for at least one of the clinical markers, so these terms can hardly be considered synonymous (despite the conveniences that would create)
The first grant I ever secured as an independent investigator was for automated digital imaging equipment for tissue microarrays, and five years later we published a massive study on its value in breast cancer. To settle the question of who is right, the machine or the human, we compared results to the gold standard of patient outcome. The bottom line is basically a tie: for nuclear immunostains, automated image analysis with appropriately sophisticated software is equal to expert visual analysis, at least in the example of breast cancer estrogen receptor expression
In a combined basic and translational study with my friend and colleague Sandi Dunn, we link the important YB-1 transcription factor to the aggressive Basal subtype of breast cancer, explaining its high rate of expression of epidermal growth factor receptor, a possible therapeutic target
A case I diagnosed as Ewing sarcoma and sent for cytogenetics is found to contain a novel translocation, fusing the unusual alternate 5' partner FUS with the rare Fifth Ewing Variant 3' partner. The result is a fusion oncogene never before described, which would have been missed by standard PCR and FISH diagnostic strategies.
This GPEC collaborative study links FOXA1, an estrogen receptor-associated transcription factor, to the good prognosis Luminal A subtype of breast cancer
In collaboration with Will Foulkes' group at McGill, we show that both triple negative and five biomarker methods for surrogate identification of the basal -like phenotype of breast cancer hold up across different institutions. Adding EGFR and CK5/6 positive markers of the basal phenotype identifies a group with particularly high risk over both short and long terms after diagnosis
This review article summarizes the state of the art in molecular pathology, local treatment by surgery and radiation, and conventional, targeted and investigational systemic therapy for sarcomas. I contributed the molecular biology and pathology sections as well as discussions on Hsp90 and histone deacetylase inhibitors. Unfortunately, the journal mislabeled the diagnoses in Figure 2 - don't worry, I can tell the difference between MFH and DSRCT!
In this followup to our 2006 PNAS article, we show that CSF1-activating translocations can be identified in about 60% of cases of tenosynovial giant cell tumor. However, the remaining cases still show high expression of CSF1, suggesting that this is, indeed, the common biological driver of this tumor, although the mechanism of activation is not always a translocation splicing this gene to a highly active promoter.
Multiple independent gene expression profiling studies of synovial sarcoma, including those from our own group and from Dr. Ladanyi's group at Memorial Sloan-Kettering, find the human groucho homolog TLE1 (a transcriptional repressor of differentiation) to be dramatically overexpressed in synovial sarcoma. We show that antibodies recognizing TLE1 are excellent nuclear biomarkers of this disease that work on clinical material with extremely high senstivity and diagnostic specificity. Primary images are available here.
With the incredible tool of a 4631-case population-based breast cancer tissue array series with long-term followup, a collaboration between the Genetic Pathology Evaluation Centre and the BCCA Breast Cancer Outcomes Unit, we are validating many exciting new biomarkers in breast cancer, with a particular focus on those identified by expression profiling studies. The first manuscript to come from this work shows the value of a new generation of diagnostic antibodies, which are superior for prognostication and prediction of tamoxifen response in breast cancer. Primary images are available here.
In this review article, we discuss the aggressive basal-like subtype of breast cancer, which tends to afflict premenopausal women, is prone to early systemic spread, and for which we currently lack effective targeted therapeutics.
Including standard as well as novel antibodies, a classification and regression tree (CART) strategy and independent tissue microarray series from Alabama, Ohio and British Columbia for validation, we helped develop an immunohistochemical panel that can assess likely patient outcomes for women with estrogen receptor positive breast cancer.
In this review article, I summarize the technologies used for microarray analysis of cancer, their advantages and limitations, and the results to date derived from applying these strategies to human primary sarcoma specimens.
Using the surrogate immunohistochemical panel we developed, this study shows that premenopausal African-American women get the more aggressive basal-like subtype of breast cancer at a frequency 2 - 3 times higher than in caucasians or in postmenopausal women of any race. This finding suggests that the poor outcomes seen in this group of women are not simply the product of differences in health care access or socioeconomic status, but rather have a major underlying biological explanation.
Following up on gene expression profiling results and in situ hybridization validation studies, we have identified a novel translocation, involving the CSF1 gene that causes the development of tenosynovial giant cell tumors and the related condition known as pigmented villonodular synovitis. A fascinating finding is that only a small fraction of the tumor cells are neoplastic, bearing the translocation; the bulk of the tumor is comprised of benign macrophages induced by the high CSF1 levels.
  • Moyano JV, Evans JR, Chen F, Lu M, Werner ME, Yehiely F, Diaz LK, Turbin D, Karaca G, Wiley E, Nielsen TO, Perou CM, Cryns VL. alphaB-Crystallin is a novel oncoprotein that predicts poor clinical outcome in breast cancer. Journal of Clinical Investigation 116:261-70 (2006).
    • Our colleague Vince Cryns, from Northwestern University, characterized an important inhibitor of breast cell apoptosis and differentiation, and proven that it is oncogenic in vitro and in mouse models. We provided clinical validation in breast cancer tissue microarrays of its approximate frequency in the population and of its association with poor prognosis and with the basal-like immunophenotype, and present primary image data on-line. The article was highlighted by an editorial and commentaries in concurrent issues of Nature and Nature Reviews Cancer.
    Using large training and validation breast cancer tissue microarrays, we demonstrate that expression of the oncoprotein MDM2 (an inhibitor of the p53 gene, the "guardian of the genome" whose function is vital in preventing malignancy) is a negative prognostic factor in breast cancer, independent of standard clinical and pathological risk factors. This paper presents the primary image data on-line.
    This was an invited minireview, based in part on presentations I had given to the Alberta Cancer Board and the Canadian Association of Pathologists, that summarizes recent advances in gene expression profiling of adult soft tissue sarcomas, and how those findings can be used to develop new targeted therapies for sarcomas.
    Our group did beta-testing and developmental work to assist our colleagues at Stanford University to develop simple programs (Excel macros) that can quickly organize large matrices of tissue microarray data. This paper describes and demonstrates how data from multiple samples of the same tumor, and multiple scores of the same sample, can be quickly and efficiently consolidated to accelerate data analysis.
  • Terry J, Lubieniecka JM, Kwan W, Liu S, Nielsen TO. Hsp90 inhibitor 17-allylamino-17-demethoxygeldanamycin prevents synovial sarcoma proliferation via apoptosis in in vitro models. Clinical Cancer Research 11:5631-8 (2005).
    •
    We show that although several specific receptor tyrosine kinases are inactive against synovial sarcoma cell line models, 17AAG, an inhibitor of multiple receptor tyrosine kinases and mutant oncoproteins, is highly effective at inducing apoptotic cell death in monolayer and spheroid growth models, at clinically-achievable concentrations. It is hoped that such findings may provide justification for a clinical trial, as effective systemic agents in this disease are currently unavailable.
    An amplicon on chromosome 8p, near the neuregulin (NRG1) gene, has been implicated by previous work in breast cancer. Using fluorescent in situ hybridization probes on our breast cancer tissue microarrays, we show in fact that the most common site of amplfication is immediately centromeric to NRG1, with the as-yet-poorly-characterized SPFH2 gene the stronger candidate oncogene, associated with poor prognosis in breast cancer.
  • Terry J, Barry TS, Hsu FD, Horsman D, Huntsman DG, Gown AM, Nielsen TO. Fluorescence in situ hybridization for the detection of t(X;18)(p11.2;q11.2) in a synovial sarcoma tissue microarray. Diagnostic Molecular Pathology 14:77-82 (2005).
    •
    We demonstrate that an SYT break-apart flourescent in-situ hybridization strategy for the detection of the translocation characteristic of synovial sarcoma is sensitive and specific on archival, formalin-fixed paraffin-embedded standard diagnostic material, even in the technically-demanding tissue microarray format, and therefore represents a readily applicable diagnostic strategy.
  • West RB, Nuyten DS, Subramanian S, Nielsen TO, Corless CL, Rubin BP, Montgomery K, Zhu S, Patel R, Hernandez-Boussard T, Goldblum JR, Brown PO, van de Vijver M, van de Rijn M. Determination of stromal signatures in breast carcinoma. PLoS Biology 3:e187 (2005).
    • The expression profiles of two non-malignant soft tissue tumors, namely desmoid-type fibromatosis and solitary fibrous tumor, are presented in detail. A fascinating surprise finding is that these same gene expression signatures are expressed in the stroma of breast cancers, and those breast cancers with the solitary fibrous tumor (epithelial support phenotype) signature have a very much worse prognosis than those with the fibromatosis (scar-like reaction phenotype) signature.
  • Wiseman SM, Makretsov N, Nielsen TO, Gilks B, Yorida E, Cheang M, Turbin D, Huntsman DG. Co-expression of the Type 1 growth factor receptor family members HER1, HER2 and HER3 has a synergistic negative prognostic effect on breast cancer survival. Cancer 103:1770-7 (2005).
    • Using tissue microarrays, we assessed by immunohistochemistry the expression of human epidermal growth factor receptor family members HER1, HER2, HER3 and HER4 and conclude that co-expression of any two of the first three is associated with especially poor prognosis.
  • Ng TL, Gown AM, Barry TS, Cheang MCU, Chan AKW, Turbin DA, Hsu FD, West RB, Nielsen TO. Nuclear beta-catenin expression in bone and soft tissue tumors. Modern Pathology 18:68-74 (2005).
    •
    Immunohistochemical analysis of four tissue microarrays representing over 500 cases of a broad variety of mesenchymal neoplasms shows that nuclear expression of beta-catenin, a transcription factor activated by Wnt signalling, is largely confined to just four tumor types. A novel digital image database is used to make primary data accessible to all readers.
  • Subramanian S, West RB, Corless CL, Ou W, Rubin BP, Chu KM, Leung SY, Yuen ST, Zhu S, Hernandez-Boussard T, Montgomery K, Nielsen TO, Patel RM, Goldblum JR, Heinrich MC, Fletcher JA, van de Rijn M. Gastrointestinal stromal tumors (GISTs) with KIT and PDGFRA mutations have distinct gene expression profiles. Oncogene 14:7780-90 (2004).
    •
    Gastrointestinal stromal tumors are shown to have distinct expression profiles, depending on whether their fundamental oncogenic mutation is in KIT exon 9, exon 11, or in PDGFRA, and this finding is validated on tissue microarrays, with implications for treatment.
  • Makretsov NA, Huntsman DG, Nielsen TO, Yorida E, Peacock M, Cheang MCU, Dunn SE, Hayes M, van de Rijn M, Bajdik C, and Gilks CB. Hierarchical clustering analysis of tissue microarray immunostaining data identifies prognostically significant groups of breast cancer. Clinical Cancer Research 10:6143-58 (2004).
    •
    Several dozen immunohistochemical markers were applied to serial sections of the same breast cancer tissue microarray. Hierarchical clustering analysis is shown to be capable of determining relationships among markers and tumors, and determining a limited panel of immunostains which are tightly linked to patient outcomes.
  • Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hernandez-Boussard T, Cowan D, Dressler L, Livasy C, Akslen LA, Ragaz J, Gown AM, Gilks CB, van de Rijn M, Perou CM.  Immunohistochemical and clinical characterization of the basal-like subtype of invasive breast carcinoma. Clinical Cancer Research 10:5367-74 (2004).
    •
    Using breast cancer tissue microarrays, we confirm the existence of the basal-type poor prognostic subgroup of breast cancer, and demonstrate its association, in many cases, with expression of the epidermal growth factor receptor, a finding with potential therapeutic implications. This paper has been particularly influential, garnering several hundred citations.
  • West RB, Harvell J, Linn SC, Liu CL, Prapong W, Montgomery K, Nielsen TO, Rubin BP, Patel R, Goldblum JR, Brown PO, van de Rijn M. Apo D in soft tissue tumors: a novel marker for dermatofibrosarcoma protuberans. American Journal of Surgical Pathology 28:1063-9 (2004).
    •
    Following up on DFSP gene expression profiling results published in Am J Pathol (Dec 2003), apolipoprotein D was identified as a specific and sensitive immunomarker for this tumor. Its value in a clinical context is here shown using tissue microarrays.
  • West RB, Corless CL, Chen X, Rubin BP, Subramanian S, Montgomery K, Zhu S, Ball CA, Nielsen TO, Patel R, Goldblum JR, Brown PO, Heinrich MC, van de Rijn M. The novel marker, DOG1, is expressed ubiquitously in GI Stromal Tumors irrespective of KIT or PDGFRA mutation status. American Journal of Pathology 165:107-13 (2004).
    Following up on the expression profiling results we published in Lancet in 2002, we identified an uncharacterized cDNA (FLJ10261), now christened DOG1 (for Discovered On GIST-1). Here we show, using sarcoma tissue microarrays, that an mRNA in situ hybridization probe, as well as a custom polyclonal antibody, created based on the cDNA spot sequence, specifically and sensitively mark gastrointestinal stromal tumors. Tumors with variant PDGF-alpha (as opposed to c-kit) activating mutations are successfully identified using this marker, which otherwise might be mis-diagnosed and inappropriately treated.
    •
  • Nielsen TO, Andrews HN, Cheang M, Kucab JE, Hsu FD, Ragaz J, Gilks CB, Makretsov N, Bajdik CD, Brookes C, Neckers LM, Evdokimova V, Huntsman DG, Dunn SE: Co-expression of the insulin-like growth factor-1 receptor (IGF-1R) and urokinase plasminogen activator (uPA) in breast cancer is associated with poor survival: Potential for intervention with 17-allylamino geldanamycin (17AAG). Cancer Research 64:286-91 (2004).
    •
    Results of gene expression profiling and array-CGH on the unusual skin tumor known as DFSP are consistent with its known translocation between chromosomes 17 and 22, and reveals a distinct set of  genes which are both amplified and overexpressed
    Initial identification, cloning and characterization of a gene (EMSY) on chromosome 11 which encodes a protein which epigenetically inactivates the BRCA2 hereditary breast and ovarian cancer gene product. EMSY is amplified in many sporadic cases of breast and ovarian cancer, and this is associated with poor prognosis. This finding helps provide an answer to the elusive question why BRCA mutations are seen in hereditary, but not the more common sporadic, types of breast cancer.

    For the month after its publication, this was the most frequently accessed article at Cell!
    A tissue microarray containing 46 cases of synovial sarcoma and 29 cases of other soft tissue tumors in its differential diagnosis is built and used to validate protein expression of two markers previously identified by cDNA microarray expression profiling. Heirarchical clustering is performed to organize immunohistochemical results, and possible therapeutic implications are discussed.
    A consensus 36 bp human replication origin DNA sequence, derived from multiple origin assay methods, is shown to drive autonomous replication in vitro and to be present at replication origin sites in human and other animal cells.
    23 standard immunohistochemical markers are tested on a tissue microarray containing over 350 examples of human tumors, and this method is found to have rapid utility for quality control evaluation of immunohistochemical staining methodologies for quality control purposes.
    A methodological paper describing the preparation of histologic slides for laser capture microdissection, the dissection procedure itself, and the isolation of RNA of sufficient quality for RT-PCR analysis from laser-captured material. An unusual case of paraganglioma, arising at the base of the tongue, is presented along with its immunohistochemical characterization. No parasympathetic ganglia are known to exist in the tongue, and this neuroendocrine tumor has never been described in this location before.
     
  • Nielsen, T.O., Cossons, N., Zannis-Hadjopoulos, M. and Price, G.B.: Circular YAC vectors containing short mammalian origin sequences are maintained under selection as HeLa episomes. Journal of Cellular Biochemistry 76:674-685 (2000).
    •  
    Human replication origin sequences cloned into a YAC vector are demonstrated to replicate autonomously in human cells. Such constructs have the potential to act as a foundation for the construction of human transfection vectors, or a defined human artificial chromosome.         The strongest arguments for legalizing euthanasia (liberty and mercy) must be balanced against legitimate fears of slippery slopes to involuntary euthanasia, erosion of palliative care, and the creation of a "duty to die." This essay spells out a detailed proposal for reconciling these concerns, involving the creation of an ethics committee to oversee legalized euthanasia in specific individual cases. Origin sequences cloned into yeast artificial chromosome vectors, as part of a project to build fully defined human artificial chromosomes, are demonstrated to associate with the nuclear matrix, a property considered necessary for proper replication origin and chromosomal function. Using a powerful, simplified version of the nascent strand PCR method for origin mapping, several autonomously replicating sequences (including examples from the anti-cruciform / mass screening experiments described in my first publication) are proven to function as in vivo chromosomal replication origins, utilized differently in normal versus transformed cell lines. This essay reviews the ethical arguments for and against germline manipulations of human embryos. The technical aspects of the procedure are discussed, as well as the medical indications for germline gene therapy. The paper concludes that the technology will always be impractical and has essentially no medical uses. A putative replication origin is identified, by sequence characteristics and an in vitro replication assay, in an imprinted chromosomal region associated with Prader-Willi and Angelman syndromes. Protein fractions are purified from human cells; those binding to an origin-containing fragment are shown to permit initiation of DNA replication in an in vitro replication system. For the first time, autonomous replication activity is demonstated in fragments derived from the well-characterized DNA replication origin located 3' to the hamster dihydrofolate-reductase gene. Fragments of mitochondrial DNA are released upon autolytic degradation of this organelle. The cytochrome oxidase subunit 3 mitochondrial gene has inserted into the c-myc genomic locus in a human cell line (HeLaTG), and interestingly is shown, in this work, to possess autonomous replication activity. Large numbers of putative human DNA replication origins are purified from genomic DNA on the basis of affinity for an anti-cruciform antibody (directed against a DNA secondary structure), followed by mass screening of clones for autonomous replication activity.

    Reprints available from: torsten@interchange.ubc.ca

    Last modified 2011.12.27

    Back to Torsten's home page.