Spatially organized cellular communities form the developing human heart

The heart, which is the first organ to develop, is highly dependent on its form to function1,2. However, how diverse cardiac cell types spatially coordinate to create the complex morphological structures that are crucial for heart function remains unclear. Here we integrated single-cell RNA-sequencing with high-resolution multiplexed error-robust fluorescence in situ hybridization to resolve the identity of the cardiac cell types that develop the human heart. This approach also provided a spatial mapping of individual cells that enables illumination of their organization into cellular communities that form distinct cardiac structures. We discovered that many of these cardiac cell types further specified into subpopulations exclusive to specific communities, which support their specialization according to the cellular ecosystem and anatomical region. In particular, ventricular cardiomyocyte subpopulations displayed an unexpected complex laminar organization across the ventricular wall and formed, with other cell subpopulations, several cellular communities. Interrogating cell-cell interactions within these communities using in vivo conditional genetic mouse models and in vitro human pluripotent stem cell systems revealed multicellular signalling pathways that orchestrate the spatial organization of cardiac cell subpopulations during ventricular wall morphogenesis. These detailed findings into the cellular social interactions and specialization of cardiac cell types constructing and remodelling the human heart offer new insights into structural heart diseases and the engineering of complex multicellular tissues for human heart repair.

Circadian modulation by time-restricted feeding rescues brain pathology and improves memory in mouse models of Alzheimer's disease

Circadian disruptions impact nearly all people with Alzheimer's disease (AD), emphasizing both their potential role in pathology and the critical need to investigate the therapeutic potential of circadian-modulating interventions. Here, we show that time-restricted feeding (TRF) without caloric restriction improved key disease components including behavioral timing, disease pathology, hippocampal transcription, and memory in two transgenic (TG) mouse models of AD. We found that TRF had the remarkable capability of simultaneously reducing amyloid deposition, increasing Aβ42 clearance, improving sleep and memory, and normalizing daily transcription patterns of multiple genes, including those associated with AD and neuroinflammation. Thus, our study unveils for the first time the pleiotropic nature of timed feeding on AD, which has far-reaching effects beyond metabolism, ameliorating neurodegeneration and the misalignment of circadian rhythmicity. Since TRF can substantially modify disease trajectory, this intervention has immediate translational potential, addressing the urgent demand for accessible approaches to reduce or halt AD progression.

Unveiling Complexity and Multipotentiality of Early Heart Fields

[Figure: see text].

A multi-tool recipe to identify regions of protein-DNA binding and their influence on associated gene expression

One commonly performed bioinformatics task is to infer functional regulation of transcription factors by observing differential expression under a knockout, and integrating DNA binding information of that transcription factor.   However, until now, this task has required dedicated bioinformatics support to perform the necessary data integration. GenomeSpace provides a protocol, or “recipe”, and a user interface with inter-operating software tools to identify protein occupancies along the genome from a ChIP-seq experiment and associated differentially regulated genes from a RNA-Seq experiment. By integrating RNA-Seq and ChIP-seq analyses, a user is easily able to associate differing expression phenotypes with changing epigenetic landscapes.

A Community Challenge for Inferring Genetic Predictors of Gene Essentialities through Analysis of a Functional Screen of Cancer Cell Lines

We report the results of a DREAM challenge designed to predict relative genetic essentialities based on a novel dataset testing 98,000 shRNAs against 149 molecularly characterized cancer cell lines. We analyzed the results of over 3,000 submissions over a period of 4 months. We found that algorithms combining essentiality data across multiple genes demonstrated increased accuracy; gene expression was the most informative molecular data type; the identity of the gene being predicted was far more important than the modeling strategy; well-predicted genes and selected molecular features showed enrichment in functional categories; and frequently selected expression features correlated with survival in primary tumors. This study establishes benchmarks for gene essentiality prediction, presents a community resource for future comparison with this benchmark, and provides insights into factors influencing the ability to predict gene essentiality from functional genetic screens. This study also demonstrates the value of releasing pre-publication data publicly to engage the community in an open research collaboration.

Modeling of RAS complexes supports roles in cancer for less studied partners

Background

RAS protein interactions have predominantly been studied in the context of the RAF and PI3kinase oncogenic pathways. Structural modeling and X-ray crystallography have demonstrated that RAS isoforms bind to canonical downstream effector proteins in these pathways using the highly conserved switch I and II regions. Other non-canonical RAS protein interactions have been experimentally identified, however it is not clear whether these proteins also interact with RAS via the switch regions.

Results

To address this question we constructed a RAS isoform-specific protein-protein interaction network and predicted 3D complexes involving RAS isoforms and interaction partners to identify the most probable interaction interfaces. The resulting models correctly captured the binding interfaces for well-studied effectors, and additionally implicated residues in the allosteric and hyper-variable regions of RAS proteins as the predominant binding site for non-canonical effectors. Several partners binding to this new interface (SRC, LGALS1, RABGEF1, CALM and RARRES3) have been implicated as important regulators of oncogenic RAS signaling. We further used these models to investigate competitive binding and multi-protein complexes compatible with RAS surface occupancy and the putative effects of somatic mutations on RAS protein interactions.

Conclusions

We discuss our findings in the context of RAS localization to the plasma membrane versus within the cytoplasm and provide a list of RAS protein interactions with possible cancer-related consequences, which could help guide future therapeutic strategies to target RAS proteins.

Electronic supplementary material

The online version of this article (doi:10.1186/s13628-017-0037-6) contains supplementary material, which is available to authorized users.

Abstract PR10: Multiple Pathway Learning accurately predicts gene essentiality in the Cancer Cell Line Encyclopedia

We applied biologically-motivated feature transformations coupled with established machine learning methods to predict gene essentiality in CCLE cell line models. By leveraging additional large datasets, such as The Cancer Genome Atlas PanCancer12 data and MSigDB pathway definitions, we improved the robustness and biological interpretability of our models. We developed a multi-pathway learning (MPL) approach that associates a genetic pathway from MSigDB with a distinct kernel for use in a multiple kernel learning setting. We evaluated the performance of MPL compared to several other regression methods including random forests, kernel ridge regression, and elastic net linear models, We combined multiple approaches using an ensemble technique on the diverse set of predictors.

We found that the best performing method was an ensemble combining MPL and random forest predictions. Both models utilized features derived from both gene expression and copy number data, the latter of which were filtered to those predicted as driver events in prior pan-cancer studies. The ensemble method was a joint winner in the recent DREAM 9 gene essentiality prediction challenge. MPL also demonstrated merit as a feature selector when used with other downstream methods.

The ensemble performed best at predicting the essentiality of genes involved in cell cycle control (cyclins and cyclin-dependent kinases), protein degradation (proteasome complex), cell proliferation signaling (sonic hedgehog, Aurora-B, RAC1), apoptosis (RB1,TP53) and hypoxia response (VEGF, VHL). Many of the key genes in those pathways are known to be drivers of cancer progression, suggesting our method's utility as a biomarker for detecting key tumorigenic events.

The advantage of MPL is that mechanistically coherent gene sets are automatically selected as high scoring pathway kernels (HSPKs). We investigated whether the HSPKs identify cellular processes relevant to the loss of key genes. To do this, we inspected the HSPKs for a few of the most abundantly mutated genes in cancer. The MPL predictor for TP53 included the targets of this transcription factor as well as HSPKs involved in apoptosis, a cellular process regulated by TP53. The retinoblastoma gene (RB1) MPL predictor included RB1 targets as well as HSPKs involved in the regulation of histone deacetylase (HDAC) that interacts with RB1 to suppress DNA synthesis. These findings suggest trends in the MPL results could reveal a pathway-level view of the synthetic lethal architecture of cells. Such a map, that links patterns of pathway expression to potential genetic vulnerabilities, could provide an invaluable tool for exploring new avenues to target cancer cells.

Citation Format: Vladislav Uzunangelov, Evan Paull, Sahil Chopra, Daniel Carlin, Adrian Bivol, Kyle Ellrott, Kiley Graim, Yulia Newton, Sam Ng, Artem Sokolov, Joshua Stuart. Multiple Pathway Learning accurately predicts gene essentiality in the Cancer Cell Line Encyclopedia. [abstract]. In: Proceedings of the AACR Special Conference on Translation of the Cancer Genome; Feb 7-9, 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 1):Abstract nr PR10.

Abstract PR02: Multiple Pathway Learning accurately predicts gene essentiality in the Cancer Cell Line Encyclopedia

We applied biologically-motivated feature transformations coupled with established machine learning methods to predict gene essentiality in CCLE cell line models. By leveraging additional large datasets, such as The Cancer Genome Atlas PanCancer12 data and MSigDB pathway definitions, we improved the robustness and biological interpretability of our models. We developed a multi-pathway learning (MPL) approach that associates a genetic pathway from MSigDB with a distinct kernel for use in a multiple kernel learning setting. We evaluated the performance of MPL compared to several other regression methods including random forests, kernel ridge regression, and elastic net linear models, We combined multiple approaches using an ensemble technique on the diverse set of predictors.

We found that the best performing method was an ensemble combining MPL and random forest predictions. Both models utilized features derived from both gene expression and copy number data, the latter of which were filtered to those predicted as driver events in prior pan-cancer studies. The ensemble method was a joint winner in the recent DREAM 9 gene essentiality prediction challenge. MPL also demonstrated merit as a feature selector when used with other downstream methods.

The ensemble performed best at predicting the essentiality of genes involved in cell cycle control (cyclins and cyclin-dependent kinases), protein degradation (proteasome complex), cell proliferation signaling (sonic hedgehog, Aurora-B, RAC1), apoptosis (RB1,TP53) and hypoxia response (VEGF, VHL). Many of the key genes in those pathways are known to be drivers of cancer progression, suggesting our method's utility as a biomarker for detecting key tumorigenic events.

The advantage of MPL is that mechanistically coherent gene sets are automatically selected as high scoring pathway kernels (HSPKs). We investigated whether the HSPKs identify cellular processes relevant to the loss of key genes. To do this, we inspected the HSPKs for a few of the most abundantly mutated genes in cancer. The MPL predictor for TP53 included the targets of this transcription factor as well as HSPKs involved in apoptosis, a cellular process regulated by TP53. The retinoblastoma gene (RB1) MPL predictor included RB1 targets as well as HSPKs involved in the regulation of histone deacetylase (HDAC) that interacts with RB1 to suppress DNA synthesis. These findings suggest trends in the MPL results could reveal a pathway-level view of the synthetic lethal architecture of cells. Such a map, that links patterns of pathway expression to potential genetic vulnerabilities, could provide an invaluable tool for exploring new avenues to target cancer cells.

Citation Format: Vladislav Uzunangelov, Evan Paull, Sahil Chopra, Daniel Carlin, Adrian Bivol, Kyle Ellrott, Kiley Graim, Yulia Newton, Sam Ng, Artem Sokolov, Joshua Stuart. Multiple Pathway Learning accurately predicts gene essentiality in the Cancer Cell Line Encyclopedia. [abstract]. In: Proceedings of the AACR Special Conference on Computational and Systems Biology of Cancer; Feb 8-11 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 2):Abstract nr PR02.

Functionally Distinct Subsets of Lineage-Biased Multipotent Progenitors Control Blood Production in Normal and Regenerative Conditions

Despite great advances in understanding the mechanisms underlying blood production, lineage specification at the level of multipotent progenitors (MPPs) remains poorly understood. Here, we show that MPP2 and MPP3 are distinct myeloid-biased MPP subsets that work together with lymphoid-primed MPP4 cells to control blood production. We find that all MPPs are produced in parallel by hematopoietic stem cells (HSCs), but with different kinetics and at variable levels depending on hematopoietic demands. We also show that the normally rare myeloid-biased MPPs are transiently overproduced by HSCs in regenerating conditions, hence supporting myeloid amplification to rebuild the hematopoietic system. This shift is accompanied by a reduction in self-renewal activity in regenerating HSCs and reprogramming of MPP4 fate toward the myeloid lineage. Our results support a dynamic model of blood development in which HSCs convey lineage specification through independent production of distinct lineage-biased MPP subsets that, in turn, support lineage expansion and differentiation.

Comprehensive, Integrative Genomic Analysis of Diffuse Lower-Grade Gliomas

BACKGROUND

Diffuse low-grade and intermediate-grade gliomas (which together make up the lower-grade gliomas, World Health Organization grades II and III) have highly variable clinical behavior that is not adequately predicted on the basis of histologic class. Some are indolent; others quickly progress to glioblastoma. The uncertainty is compounded by interobserver variability in histologic diagnosis. Mutations in IDH, TP53, and ATRX and codeletion of chromosome arms 1p and 19q (1p/19q codeletion) have been implicated as clinically relevant markers of lower-grade gliomas.

METHODS

We performed genomewide analyses of 293 lower-grade gliomas from adults, incorporating exome sequence, DNA copy number, DNA methylation, messenger RNA expression, microRNA expression, and targeted protein expression. These data were integrated and tested for correlation with clinical outcomes.

RESULTS

Unsupervised clustering of mutations and data from RNA, DNA-copy-number, and DNA-methylation platforms uncovered concordant classification of three robust, nonoverlapping, prognostically significant subtypes of lower-grade glioma that were captured more accurately by IDH, 1p/19q, and TP53 status than by histologic class. Patients who had lower-grade gliomas with an IDH mutation and 1p/19q codeletion had the most favorable clinical outcomes. Their gliomas harbored mutations in CIC, FUBP1, NOTCH1, and the TERT promoter. Nearly all lower-grade gliomas with IDH mutations and no 1p/19q codeletion had mutations in TP53 (94%) and ATRX inactivation (86%). The large majority of lower-grade gliomas without an IDH mutation had genomic aberrations and clinical behavior strikingly similar to those found in primary glioblastoma.

CONCLUSIONS

The integration of genomewide data from multiple platforms delineated three molecular classes of lower-grade gliomas that were more concordant with IDH, 1p/19q, and TP53 status than with histologic class. Lower-grade gliomas with an IDH mutation either had 1p/19q codeletion or carried a TP53 mutation. Most lower-grade gliomas without an IDH mutation were molecularly and clinically similar to glioblastoma. (Funded by the National Institutes of Health.).

Ethical and legal aspects of global tobacco control

On 28 February 2005, the Framework Convention on Tobacco Control came into force as a result of at least 40 countries becoming State Parties through ratification of this first ever health treaty sponsored by the World Health Organization. This article discusses the bioethical, trade, and legal aspects of global tobacco control. Special emphasis is given to globalisation of tobacco use and the challenges it poses to sovereign nations. It also advocates a bioethical basis in the pursuit of global solutions to expanding tobacco use.

Multicenter US study of hospital resource utilization associated with cytomegalovirus-related readmission of renal and heart transplant patients

Incidence of cytomegalovirus (CMV)-related rehospitalization and associated resource use were captured by the Transplant Infection Cost Analysis (TICA) program, which examined patient records and hospital billing data in multiple solid organ transplant centers in the US. The experiences of two adult heart and three adult renal transplant centers were each pooled for analysis. Financial data were standardized to 1998 US dollars using the Medical Care component of the US Consumer Price Index. CMV-related readmissions among renal transplant patients averaged 10.5 days (range 1-56) with average charges of $22,598. Heart transplant patients readmitted for CMV incurred an average charge of $42,111 and average hospital stay of 10.9 days (range 2-95). CMV-related hospital resource use represented a significant portion of the average cost of the original transplant and associated length of stay.

A double-blind placebo-controlled randomized phase III trial of 5-fluorouracil and leucovorin, plus or minus trimetrexate, in previously untreated patients with advanced colorectal cancer

BACKGROUND

Trimetrexate (TMTX) biochemically modulates 5-fluorouracil (5-FU) and leucovorin (LCV). Two phase II trials demonstrated promising activity for TMTX/5-FU/LCV in patients with untreated advanced colorectal cancer (ACC). This trial was designed to demonstrate the safety and efficacy of TMTX/5-FU/LCV as first-line treatment in ACC.

PATIENTS AND METHODS

Eligible patients with ACC were randomized in double-blind fashion to receive placebo or TMTX (110 mg/m2) intravenously (i.v.) followed 24 h later by i.v. LCV 200 mg/m2, and 5-FU 500 mg/m2 plus oral LCV rescue. Both schedules were given weekly for 6 weeks every 8 weeks. Patients were evaluated for progression-free survival (PFS), overall survival (OS), tumor response, quality of life (QoL) and toxicity.

RESULTS

A total of 382 eligible patients were randomized. Significant toxicities were noted more frequently with TMTX/5-FU/LCV. Diarrhea was the most common grade 3 or 4 side-effect (41% and 28% on the TMTX and placebo arms, respectively). QoL scores and response rates did not differ between treatment arms. PFS was 5.3 months and 4.4 months in the TMTX and placebo arms, respectively (P = 0.77; Wilcoxon). OS was 15.8 months and 16.8 months, respectively (P = 0.73; Wilcoxon).

CONCLUSIONS

The addition of TMTX to a weekly regimen of 5-FU/LCV worsened grade 3 or 4 diarrhea. The inclusion of TMTX did not yield any significant improvements in response rate, PFS or OS.

ROC surface: a generalization of ROC curve analysis

Receiver operating characteristic (ROC) curve analysis is widely used in biomedical research to assess the performance of diagnostic tests. Much of the work has been directed at developing accurate indices to describe ROC curves and appropriate statistics to test differences between them. The analysis, however, is largely built on the assumption that the test results are dichotomous. We generalize the ROC curve analysis to allow for tests to have more than two outcomes. The generalized ROC curve constitutes a surface. We propose to use the volume under the surface to measure the accuracy of a diagnostic test.

Planning impairments in frontal lobe dementia and frontal lobe lesion patients

Patients with frontal lobe brain damage are reportedly impaired on tasks that require plan development and execution. In this study, we examined the performance of 15 patients diagnosed with frontal lobe dementia and 14 patients with focal frontal lobe lesions on the Tower of London planning task. Patients with frontal lobe dementia committed a significantly higher number of rule violations, made more moves, and demonstrated longer solution time latencies compared to their matched controls. Patients with frontal lobe lesions demonstrated significantly delayed solution times and also made more moves compared to their matched controls. Frontal lobe lesion patient performance suggests an impairment in execution-related processes, while frontal lobe dementia patients appear to be impaired in both plan development and execution. Despite these findings, the identification of a specific cognitive impairment that induces these planning problems remains elusive.

Lumbar disc high-intensity zone: the value and significance of provocative discography in the determination of the discogenic pain source

Disagreement still exists in the literature as to the significance of the high-intensity zone (HIZ) demonstrated on magnetic resonance imaging (MRI) as a potential pain indicator in patients with low back pain. A prospective blind study was therefore conducted to evaluate the lumbar disc high-intensity zone with the pain provocation response of lumbar discography. Consecutive patients with low back pain unresponsive to conservative treatment and being considered for spinal fusion were subjected to MRI followed by lumbar discography as a pre-operative assessment. The discographer was blinded to the results of the MRI scans. We used the chi-squared test to analyse our results. Ninety-two HIZs were identified in 73 patients, mainly occurring at L4/5 (48%) and L5/S1 (35%). Significant correlation was found between abnormal disc morphology and the HIZ (P < 0.001). In morphologically abnormal discs (grades 3, 4 and 5), there was a significant correlation between the HIZ and exact or similar pain reproduction (P < 0.001). The sensitivity, specificity and positive predictive value for pain reproduction were high, at 81%, 79% and 87% respectively. The nature of the HIZ remains unknown, but it may represent an area of secondary inflammation as a result of an annular tear. We conclude from our study that the lumbar disc HIZ observed on MRI in patients with low back pain is likely to represent painful internal disc disruption.

A direct comparison of the activities of two humanized respiratory syncytial virus monoclonal antibodies: MEDI-493 and RSHZl9

Two humanized monoclonal antibodies, MEDI-493 and RSHZ19, were developed independently as potential improvements over RSV-IGIV for prevention of respiratory syncytial virus (RSV) infection. RSV-IGIV is a polyclonal human antibody preparation for intravenous infusion enriched for RSV neutralizing activity. A phase III clinical trial showed that MEDI-493 significantly reduced hospitalizations due to RSV infection. In a separate trial, RSHZ19 failed to show significant efficacy. In new studies, the in vitro and in vivo activities of MEDI-493 and RSHZ19 were compared to determine whether the different clinical results are related to differences in biologic activity. MEDI-493 was consistently 4- to 5-fold more potent than RSHZ19 in antigen binding, RSV neutralization, and fusion inhibition assays. Although both MEDI-493 and RSHZ19 were effective against A and B subtypes of RSV in the cotton rat model of RSV infection, 2- to 4-fold higher doses of RSHZ19 were required for similar protection. The enhanced activity of MEDI-493 compared with RSHZ19 may, in part, explain its better clinical effect.

Reduction of respiratory syncytial virus (RSV) in tracheal aspirates in intubated infants by use of humanized monoclonal antibody to RSV F protein

Thirty-five children <2 years of age mechanically ventilated for respiratory syncytial virus (RSV) infection were randomized to receive an intravenous infusion of 15 mg/kg MEDI-493 or placebo. RSV concentration was measured in tracheal secretions by plaque assay before and at 24-h intervals after treatment. The reduction in tracheal RSV concentration from day 0 to day 1 (-1.7+/-0.28 vs. -0. 6+/-0.21 log10 pfu/mL; P=.004) and from day 0 to day 2 (-2.5+/-0.26 vs. -1.0+/-0.41 log10 pfu/mL; P=.012) was significantly greater in the MEDI-493 group than in the placebo group. RSV concentration in nasal aspirates did not differ significantly between the groups. No significant differences were observed in the tracheal aspirate white blood cell count, or myeloperoxidase or eosinophilic cationic protein concentration, or in measures of disease severity between the groups. Thus, treatment with 15 mg/kg MEDI-493 intravenously was well-tolerated and significantly reduced RSV concentration in tracheal aspirates of children with respiratory failure due to RSV.

Safety, tolerance and pharmacokinetics of a humanized monoclonal antibody to respiratory syncytial virus in premature infants and infants with bronchopulmonary dysplasia. MEDI-493 Study Group

BACKGROUND

Respiratory syncytial virus (RSV) is the most common cause of lower respiratory tract infection in infants. MEDI-493 (palivizumab) is a humanized monoclonal antibody to the fusion protein of RSV and is active in animal models for prevention of pulmonary RSV replication.

OBJECTIVE

To describe the safety, tolerance, immunogenicity and pharmacokinetics of repeat intravenous doses of MEDI-493 in premature infants or infants with bronchopulmonary dysplasia.

DESIGN

Phase I/II multicenter, randomized, double blind, placebo-controlled, dose escalation trial.

PATIENT POPULATION

Infants born prematurely (< or = 35 weeks of gestation) who were < or = 6 months of age and infants with bronchopulmonary dysplasia who were < or = 24 months of age were eligible for study participation. STUDY AGENTS: Participants received 3, 10 or 15 mg/kg MEDI-493 or 0.9% saline intravenously every 30 days for up to five doses.

RESULTS

MEDI-493 was safe and well-tolerated and did not induce a specific anti-MEDI-493 response. The mean half-life of 20 days was comparable with that of other immunoglobulin G preparations. Mean trough serum concentrations 30 days after Infusion 1 were 6.8, 36.1 and 60.6 microg/ml for the 3-, 10- and 15-mg/kg dose groups, respectively. After Infusion 2 the trough concentrations were 11.9, 45.2 and 70.7 microg/ml. After subsequent doses the mean trough values ranged from 14 to 18 microg/ml in those given 3 mg/kg and were > 40 microg/ml for patients who received 10 or 15 mg/kg MEDI-493 (46 to 72 microg/ml and 88 to 96 microg/ml, respectively).

CONCLUSIONS

MEDI-493 was safe and well-tolerated in this high risk pediatric population. Mean serum concentrations of MEDI-493 that have been shown to produce a 2-log reduction in pulmonary RSV titer in cotton rats were maintained when 10 or 15 mg/kg MEDI-493 was given every 30 days to pediatric patients at high risk for serious RSV disease. Monthly doses of 15 mg/kg maintained concentrations of > 40 microg/ml for the majority of patients.

CA15-3 serum levels in breast cancer and other malignancies--correlation with clinical course

Serum levels of the breast cancer-associated tumor marker CA15-3 were evaluated in three patient groups: breast, colorectal and ovarian cancer and in healthy subjects. Of 51 blood samples obtained from 31 patients with metastatic breast cancer (Stage IV disease), 98% had marker levels greater than 30 u/ml and 86% had levels greater than 50 u/ml. In contrast, of 49 samples from 42 patients with Stage I-II disease, 45% had levels greater than 30 u/ml but only 6% had levels greater than 50 u/ml (mean 29.5 +/- 18 u/ml). The mean level of the CA15-3 antigen in patients with Stage IV breast cancer and responding to therapy was 79.8 +/- 27 u/ml, while the mean level in patients not responding to therapy was 134 +/- 66 u/ml (P less than 0.02). The mean serial changes in CA15-3 levels for those responding to therapy was -28.4% while the mean change for those not responding to therapy was +44%. The mean marker level for 26 patients with colorectal carcinoma was 29.8 +/- 29 u/ml; 23% of these patients had levels greater than 30 u/ml and 7% had levels greater than 50 u/ml. No substantial difference was seen in those with active compared with nonactive colorectal carcinoma. The mean marker level for 14 patients with active ovarian carcinoma was 83 +/- 62 u/ml. Of these patients, 78% had CA15-3 levels greater than 30 u/ml and 50% had levels greater than 50 u/ml. All healthy subjects (n = 22) had marker levels less than 30 u/ml. We compared CA15-3 and CEA blood levels in the same patient population; 86% of patients with metastatic breast cancer (Stage IV disease) had CA15-3 levels greater than 50 u/ml while only 72% of these patients had CEA levels greater than 5 ng/ml. These findings suggest that the CA15-3 assay reflects the clinical course of patients with advanced breast cancer and may be superior to CEA as a monitor of therapeutic efficacy.