LMD: Multiscale Marker Identification in Single-cell RNA-seq Data.

Accurate cell marker identification in single-cell RNA-seq data is crucial for understanding cellular diversity and function. An ideal marker is highly specific in identifying cells that are similar in terms of function and state. Current marker identification methods, commonly based on clustering and differential expression, capture general cell-type markers but often miss markers for subtypes or functional cell subsets, with their performance largely dependent on clustering quality. Moreover, cluster-independent approaches tend to favor genes that lack the specificity required to characterize regions within the transcriptomic space at multiple scales. Here we introduce Localized Marker Detector (LMD), a novel tool to identify "localized genes" - genes with expression profiles specific to certain groups of highly similar cells - thereby characterizing cellular diversity in a multi-resolution and fine-grained manner. LMD's strategy involves building a cell-cell affinity graph, diffusing the gene expression value across the cell graph, and assigning a score to each gene based on its diffusion dynamics. We show that LMD exhibits superior accuracy in recovering known cell-type markers in the Tabula Muris bone marrow dataset relative to other methods for marker identification. Notably, markers favored by LMD exhibit localized expression, whereas markers prioritized by other clustering-free algorithms are often dispersed in the transcriptomic space. We further group the markers suggested by LMD into functional gene modules to improve the separation of cell types and subtypes in a more fine-grained manner. These modules also identify other sources of variation, such as cell cycle status. In conclusion, LMD is a novel algorithm that can identify fine-grained markers for cell subtypes or functional states without relying on clustering or differential expression analysis. LMD exploits the complex interactions among cells and reveals cellular diversity at high resolution.

Immune dysfunction revealed by digital spatial profiling of immuno-oncology markers in progressive stages of renal cell carcinoma and in brain metastases.

BACKGROUND: The tumor microenvironment (TME) contributes to cancer progression and treatment response to therapy, including in renal cell carcinoma (RCC). Prior profiling studies, including single-cell transcriptomics, often involve limited sample sizes and lack spatial orientation. The TME of RCC brain metastases, a major cause of morbidity, also remains largely uncharacterized. METHODS: We performed digital spatial profiling on the NanoString GeoMx platform using 52 validated immuno-oncology markers on RCC tissue microarrays representing progressive stages of RCC, including brain metastases. We profiled 76 primary tumors, 27 adjacent histologically normal kidney samples, and 86 metastases, including 24 brain metastases. RESULTS: We observed lower immune checkpoint (TIM-3 and CTLA-4), cytolytic (GZMA and GZMB), and T cell activation (CD25) protein expression in metastases compared with primary tumors in two separate cohorts. We also identified changes in macrophages in metastases, with brain metastases-susceptible patients showing less M1-like, inflammatory macrophage markers (HLA-DR and CD127) in metastatic samples. A comparison of brain metastases to extracranial metastases revealed higher expression of the anti-apoptotic, BCL-2-family protein BCL-XL and lower expression of the innate immune activator STING in brain metastases. Lower TIM-3 and CD40 in the TME of brain metastases appear to be associated with longer survival, a finding that requires further validation. CONCLUSIONS: Compared with primary tumors, RCC metastases, including brain metastases, express lower levels of numerous markers of immune activation and current or investigational therapeutic targets. Our findings may have important implications for designing future biomarker and treatment studies and may aid in development of brain metastases-specific therapies.

Impact of healthcare worker shift scheduling on workforce preservation during the COVID-19 pandemic.

Reducing severe acute respiratory coronavirus virus 2 (SARS-CoV-2) infections among healthcare workers is critical. We ran Monte Carlo simulations modeling the spread of SARS-CoV-2 in non-COVID-19 wards, and we found that longer nursing shifts and scheduling designs in which teams of nurses and doctors co-rotate no more frequently than every 3 days can lead to fewer infections.

A multicriteria approach for risk assessment of Covid-19 in urban district lockdown.

At the beginning of 2020, the spread of a new strand of Coronavirus named SARS-CoV-2 (COVID-19) raised the interest of the scientific community about the risk assessment related to the viral infection. The contagion became pandemic in few months forcing many Countries to declare lockdown status. In this context of quarantine, all commercial and productive activities are suspended, and many Countries are experiencing a serious crisis. To this aim, the understanding of risk of contagion in every urban district is fundamental for governments and administrations to establish reopening strategies. This paper proposes the calibration of an index able to predict the risk of contagion in urban districts in order to support the administrations in identifying the best strategies to reduce or restart the local activities during lockdown conditions. The objective regards the achievement of a useful tool to predict the risk of contagion by considering socio-economic data such as the presence of activities, companies, institutions and number of infections in urban districts. The proposed index is based on a factorial formula, simple and easy to be applied by practitioners, calibrated by using an optimization-based procedure and exploiting data of 257 urban districts of Apulian region (Italy). Moreover, a comparison with a more refined analysis, based on the training of Artificial Neural Networks, is performed in order to take into account the non-linearity of the phenomenon. The investigation quantifies the influence of each considered parameter in the risk of contagion useful to obtain risk analysis and forecast scenarios.

MET Expression in Primary and Metastatic Clear Cell Renal Cell Carcinoma: Implications of Correlative Biomarker Assessment to MET Pathway Inhibitors.

AIMS: Inhibitors of the MET pathway hold promise in the treatment for metastatic kidney cancer. Assessment of predictive biomarkers may be necessary for appropriate patient selection. Understanding MET expression in metastases and the correlation to the primary site is important, as distant tissue is not always available. METHODS AND RESULTS: MET immunofluorescence was performed using automated quantitative analysis and a tissue microarray containing matched nephrectomy and distant metastatic sites from 34 patients with clear cell renal cell carcinoma. Correlations between MET expressions in matched primary and metastatic sites and the extent of heterogeneity were calculated. The mean expression of MET was not significantly different between primary tumors when compared to metastases (P = 0.1). MET expression weakly correlated between primary and matched metastatic sites (R = 0.5) and a number of cases exhibited very high levels of discordance between these tumors. Heterogeneity within nephrectomy specimens compared to the paired metastatic tissues was not significantly different (P = 0.39). CONCLUSIONS: We found that MET expression is not significantly different in primary tumors than metastatic sites and only weakly correlates between matched sites. Moderate concordance of MET expression and significant expression heterogeneity may be a barrier to the development of predictive biomarkers using MET targeting agents.

Regulation of glutamine carrier proteins by RNF5 determines breast cancer response to ER stress-inducing chemotherapies.

Many tumor cells are fueled by altered metabolism and increased glutamine (Gln) dependence. We identify regulation of the L-glutamine carrier proteins SLC1A5 and SLC38A2 (SLC1A5/38A2) by the ubiquitin ligase RNF5. Paclitaxel-induced ER stress to breast cancer (BCa) cells promotes RNF5 association, ubiquitination, and degradation of SLC1A5/38A2. This decreases Gln uptake, levels of TCA cycle components, mTOR signaling, and proliferation while increasing autophagy and cell death. Rnf5-deficient MMTV-PyMT mammary tumors were less differentiated and showed elevated SLC1A5 expression. Whereas RNF5 depletion in MDA-MB-231 cells promoted tumorigenesis and abolished paclitaxel responsiveness, SLC1A5/38A2 knockdown elicited opposing effects. Inverse RNF5(hi)/SLC1A5/38A2(lo) expression was associated with positive prognosis in BCa. Thus, RNF5 control of Gln uptake underlies BCa response to chemotherapies.

Preanalytical variables and phosphoepitope expression in FFPE tissue: quantitative epitope assessment after variable cold ischemic time.

Individualized targeted therapies for cancer patients require accurate and reproducible assessment of biomarkers to be able to plan treatment accordingly. Recent studies have shown highly variable effects of preanalytical variables on gene expression profiling and protein levels of different tissue types. Several publications have described protein degradation of tissue samples as a direct result of delay of formalin fixation of the tissue. Phosphorylated proteins are more labile and epitope degradation can happen within 30 min of cold ischemic time. To address this issue, we evaluated the change in antigenicity of a series of phosphoproteins in paraffin-embedded samples from breast tumors as a function of time to formalin fixation. A tissue microarray consisting of 93 breast cancer specimens with documented time-to-fixation was used to evaluate changes in antigenicity of 12 phosphoepitopes frequently used in research settings as a function of cold ischemic time. Analysis was performed in a quantitative manner using the AQUA technology for quantitative immunofluorescence. For each marker, least squares univariate linear regression was performed and confidence intervals were computed using bootstrapping. The majority of the epitopes tested revealed changes in expression levels with increasing time to formalin fixation. Some phosphorylated proteins, such as phospho-HSP27 and phospho-S6 RP, involved in post-translational modification and stress response pathways increased in expression or phosphorylation levels. Others (like phospho-AKT, phosphor-ERK1/2, phospho-Tyrosine, phospho-MET, and others) are quite labile and loss of antigenicity can be reported within 1-2 h of cold ischemic time. Therefore specimen collection should be closely monitored and subjected to quality control measures to ensure accurate measurement of these epitopes. However, a few phosphoepitopes (like phospho-JAK2 and phospho-ER) are sufficiently robust for routine usage in companion diagnostic testing.

PLEKHA5 as a Biomarker and Potential Mediator of Melanoma Brain Metastasis.

PURPOSE: Approximately 40% of patients with metastatic melanoma develop brain metastases. Our purpose was to identify genes aberrantly expressed in melanoma that might be associated with propensity for brain homing. EXPERIMENTAL DESIGN: We studied gene expression profiles in a cell line model of brain metastasis (cerebrotropic A375Br cells vs. parental A375P cells) and compared them with profiles of patients who developed early brain metastases and who did not. A tissue microarray containing 169 metastatic melanoma cases with variable time to brain metastasis was constructed to further study marker expression by quantitative immunofluorescence. An in vitro model of the blood brain barrier (BBB) was generated to evaluate potential mediators of brain metastases. RESULTS: PLEKHA5 was differentially expressed in both the A375 cell line model and patient samples subjected to gene expression profiling. At the protein level, by quantitative immunofluorescence, PLEKHA5 was associated with decreased brain metastasis-free survival. PLEKHA5 overexpression was not associated with other metastatic sites. Knockdown of PLEKHA5 decreases the viability of A375Br cells, inhibits BBB transmigration and invasion in vitro. Similar results were found with YUMUL cells, cultured from a patient with overwhelming brain metastases. PLEKHA5 knockdown did not affect the viability of A375P cells. CONCLUSIONS: PLEKHA5 expression in melanoma tumors was associated with early development of brain metastases. Inhibition of PLEKHA5 might decrease passage across the BBB and decrease proliferation and survival of melanoma cells both in the brain and in extracerebral sites.

Interactions with RNA direct the Polycomb group protein SCML2 to chromatin where it represses target genes.

Polycomb repressive complex-1 (PRC1) is essential for the epigenetic regulation of gene expression. SCML2 is a mammalian homolog of Drosophila SCM, a Polycomb-group protein that associates with PRC1. In this study, we show that SCML2A, an SCML2 isoform tightly associated to chromatin, contributes to PRC1 localization and also directly enforces repression of certain Polycomb target genes. SCML2A binds to PRC1 via its SPM domain and interacts with ncRNAs through a novel RNA-binding region (RBR). Targeting of SCML2A to chromatin involves the coordinated action of the MBT domains, RNA binding, and interaction with PRC1 through the SPM domain. Deletion of the RBR reduces the occupancy of SCML2A at target genes and overexpression of a mutant SCML2A lacking the RBR causes defects in PRC1 recruitment. These observations point to a role for ncRNAs in regulating SCML2 function and suggest that SCML2 participates in the epigenetic control of transcription directly and in cooperation with PRC1.DOI: http://dx.doi.org/10.7554/eLife.02637.001.

NY-ESO-1 as a potential immunotherapeutic target in renal cell carcinoma.

BACKGROUND: Novel immune therapies targeting tumor specific antigens are being developed. Our purpose was to determine expression of the cancer testes antigen NY-ESO-1 in renal cell carcinoma (RCC), as NY-ESO-1 targeting approaches, particularly adoptive cell therapy, have not been evaluated in this disease. METHODS: We employed tissue microarrays containing >300 unique RCC cases and adjacent benign renal tissue to determine NY-ESO-1 expression using a quantitative immunofluorescence method. In addition, we studied NY-ESO-1 expression in 35 matched primary and metastatic RCC specimens to assess concordance between different tumor sites. RESULTS: NY-ESO-1 was highly expressed in a subset of RCCs. Expression in primary RCC specimens was significantly higher than adjacent normal renal tissue (P<0.0001) and higher in clear cell carcinomas than papillary RCC (P<0.0001). Expression levels in metastatic specimens were higher than in matched primary samples (P=0.0018), and the correlation between the two sites was modest (χ2=3.5, p=0.06). CONCLUSIONS: Aberrant NY-ESO-1 expression seen in clear cell RCC suggests that NY-ESO-1 targeting approaches should be studied in this disease. Expression is higher in metastatic sites, and discordance between primary and metastatic sites in some patients suggests that patient selection for these therapies should be based on expression in metastatic rather than nephrectomy specimens.

Oct-1 regulates IL-17 expression by directing interchromosomal associations in conjunction with CTCF in T cells.

Interchromosomal associations can regulate gene expression, but little is known about the molecular basis of such associations. In response to antigen stimulation, naive T cells can differentiate into Th1, Th2, and Th17 cells expressing IFN-γ, IL-4, and IL-17, respectively. We previously reported that in naive T cells, the IFN-γ locus is associated with the Th2 cytokine locus. Here we show that the Th2 locus additionally associates with the IL-17 locus. This association requires a DNase I hypersensitive region (RHS6) at the Th2 locus. RHS6 and the IL-17 promoter both bear Oct-1 binding sites. Deletion of either of these sites or Oct-1 gene impairs the association. Oct-1 and CTCF bind their cognate sites cooperatively, and CTCF deficiency similarly impairs the association. Finally, defects in the association lead to enhanced IL-17 induction. Collectively, our data indicate Th17 lineage differentiation is restrained by the Th2 locus via interchromosomal associations organized by Oct-1 and CTCF.

A tissue quality index: an intrinsic control for measurement of effects of preanalytical variables on FFPE tissue.

While efforts are made to improve tissue quality and control preanalytical variables, pathologists are often confronted with the challenge of molecular analysis of patient samples of unknown quality. Here we describe a first attempt to construct a tissue quality index (TQI) or an intrinsic control that would allow a global assessment of protein status based on quantitative measurement of a small number of selected, informative epitopes. Quantitative immunofluorescence (QIF) of a number of proteins was performed on a series of 93 breast cancer cases where levels of expression were assessed as a function of delayed time to formalin fixation. A TQI was constructed based on the combination of proteins that most accurately reflect increased and decreased levels of expression in proportion to delay time. The TQI, defined by combinations of measurements of cytokeratin, ERK1/2 and pHSP-27 and their relationship to cold ischemic time were validated on a second build of the training series and on two independent breast tissue cohorts with recorded time to formalin fixation. We show an association of negative TQI values (an indicator for loss of tissue quality) with increasing cold ischemic time on both validation cohorts and an association with loss of ER expression levels on all three breast cohorts. Using expression levels of three epitopes, we can begin to assess the likelihood of delayed time to fixation or decreased tissue quality. This TQI represents a proof of concept for the use of epitope expression to provide a mechanism for monitoring tissue quality.

Ranking and combining multiple predictors without labeled data.

In a broad range of classification and decision-making problems, one is given the advice or predictions of several classifiers, of unknown reliability, over multiple questions or queries. This scenario is different from the standard supervised setting, where each classifier's accuracy can be assessed using available labeled data, and raises two questions: Given only the predictions of several classifiers over a large set of unlabeled test data, is it possible to (i) reliably rank them and (ii) construct a metaclassifier more accurate than most classifiers in the ensemble? Here we present a spectral approach to address these questions. First, assuming conditional independence between classifiers, we show that the off-diagonal entries of their covariance matrix correspond to a rank-one matrix. Moreover, the classifiers can be ranked using the leading eigenvector of this covariance matrix, because its entries are proportional to their balanced accuracies. Second, via a linear approximation to the maximum likelihood estimator, we derive the Spectral Meta-Learner (SML), an unsupervised ensemble classifier whose weights are equal to these eigenvector entries. On both simulated and real data, SML typically achieves a higher accuracy than most classifiers in the ensemble and can provide a better starting point than majority voting for estimating the maximum likelihood solution. Furthermore, SML is robust to the presence of small malicious groups of classifiers designed to veer the ensemble prediction away from the (unknown) ground truth.

TrAp: a tree approach for fingerprinting subclonal tumor composition.

Revealing the clonal composition of a single tumor is essential for identifying cell subpopulations with metastatic potential in primary tumors or with resistance to therapies in metastatic tumors. Sequencing technologies provide only an overview of the aggregate of numerous cells. Computational approaches to de-mix a collective signal composed of the aberrations of a mixed cell population of a tumor sample into its individual components are not available. We propose an evolutionary framework for deconvolving data from a single genome-wide experiment to infer the composition, abundance and evolutionary paths of the underlying cell subpopulations of a tumor. We have developed an algorithm (TrAp) for solving this mixture problem. In silico analyses show that TrAp correctly deconvolves mixed subpopulations when the number of subpopulations and the measurement errors are moderate. We demonstrate the applicability of the method using tumor karyotypes and somatic hypermutation data sets. We applied TrAp to Exome-Seq experiment of a renal cell carcinoma tumor sample and compared the mutational profile of the inferred subpopulations to the mutational profiles of single cells of the same tumor. Finally, we deconvolve sequencing data from eight acute myeloid leukemia patients and three distinct metastases of one melanoma patient to exhibit the evolutionary relationships of their subpopulations.

Arpeggio: harmonic compression of ChIP-seq data reveals protein-chromatin interaction signatures.

Researchers generating new genome-wide data in an exploratory sequencing study can gain biological insights by comparing their data with well-annotated data sets possessing similar genomic patterns. Data compression techniques are needed for efficient comparisons of a new genomic experiment with large repositories of publicly available profiles. Furthermore, data representations that allow comparisons of genomic signals from different platforms and across species enhance our ability to leverage these large repositories. Here, we present a signal processing approach that characterizes protein-chromatin interaction patterns at length scales of several kilobases. This allows us to efficiently compare numerous chromatin-immunoprecipitation sequencing (ChIP-seq) data sets consisting of many types of DNA-binding proteins collected from a variety of cells, conditions and organisms. Importantly, these interaction patterns broadly reflect the biological properties of the binding events. To generate these profiles, termed Arpeggio profiles, we applied harmonic deconvolution techniques to the autocorrelation profiles of the ChIP-seq signals. We used 806 publicly available ChIP-seq experiments and showed that Arpeggio profiles with similar spectral densities shared biological properties. Arpeggio profiles of ChIP-seq data sets revealed characteristics that are not easily detected by standard peak finders. They also allowed us to relate sequencing data sets from different genomes, experimental platforms and protocols. Arpeggio is freely available at http://sourceforge.net/p/arpeggio/wiki/Home/.

Expression of drug targets in primary and matched metastatic renal cell carcinoma tumors.

BACKGROUND: Targeted therapies in renal cell carcinoma can have different effects on primary and metastatic tumors. To pave the way for predictive biomarker development, we assessed differences in expression of targets of currently approved drugs in matched primary and metastatic specimens from 34 patients. METHODS: Four cores from each site were embedded in tissue microarray blocks. Expression of B-Raf, C-Raf, cKIT, FGF-R1, HIF-2α, mTOR, PDGF-Rß, VEGF-R1, VEGF-R2, VEGF-R3, VEGF, VEGF-B, VEGF-C, VEGF-D, MEK1, and ERK1/2 was studied using a quantitative immunofluorescence method. RESULTS: No significant differences were observed in global expression levels in primary and metastatic renal cell carcinoma tumors, with the exception of MEK, which had higher expression in metastatic than primary specimens. Similarly, more ki67 positive cells were seen in metastatic specimens. Correlations between marker expression in primary and metastatic specimens were variable, with the lowest correlation seen for FGF-R1 and VEGF-D. There were no significant differences in the degree of heterogeneity in primary versus metastatic tumors. CONCLUSIONS: Expression of most of the studied markers was similar in primary and metastatic renal cell carcinoma tumors, suggesting that predictive biomarker testing for these markers can be conducted on either the primary or metastatic tumors for most markers.

Quantitative assessment of effect of preanalytic cold ischemic time on protein expression in breast cancer tissues.

BACKGROUND: Companion diagnostic tests can depend on accurate measurement of protein expression in tissues. Preanalytic variables, especially cold ischemic time (time from tissue removal to fixation in formalin) can affect the measurement and may cause false-negative results. We examined 23 proteins, including four commonly used breast cancer biomarker proteins, to quantify their sensitivity to cold ischemia in breast cancer tissues. METHODS: A series of 93 breast cancer specimens with known time-to-fixation represented in a tissue microarray and a second series of 25 matched pairs of core needle biopsies and breast cancer resections were used to evaluate changes in antigenicity as a function of cold ischemic time. Estrogen receptor (ER), progesterone receptor (PgR), HER2 or Ki67, and 19 other antigens were tested. Each antigen was measured using the AQUA method of quantitative immunofluorescence on at least one series. All statistical tests were two-sided. RESULTS: We found no evidence for loss of antigenicity with time-to-fixation for ER, PgR, HER2, or Ki67 in a 4-hour time window. However, with a bootstrapping analysis, we observed a trend toward loss for ER and PgR, a statistically significant loss of antigenicity for phosphorylated tyrosine (P = .0048), and trends toward loss for other proteins. There was evidence of increased antigenicity in acetylated lysine, AKAP13 (P = .009), and HIF1A (P = .046), which are proteins known to be expressed in conditions of hypoxia. The loss of antigenicity for phosphorylated tyrosine and increase in expression of AKAP13, and HIF1A were confirmed in the biopsy/resection series. CONCLUSIONS: Key breast cancer biomarkers show no evidence of loss of antigenicity, although this dataset assesses the relatively short time beyond the 1-hour limit in recent guidelines. Other proteins show changes in antigenicity in both directions. Future studies that extend the time range and normalize for heterogeneity will provide more comprehensive information on preanalytic variation due to cold ischemic time.

Integrated analysis of tumor samples sheds light on tumor heterogeneity.

The heterogeneity of tumor samples is a major challenge in the analysis of high-throughput profiling of tumor biopsies and cell lines. The measured aggregate signals of multigenerational progenies often represent an average of several tumor subclones with varying genomic aberrations and different gene expression levels. The goal of the present study was to integrate copy number analyses from SNP-arrays and karyotyping, gene expression profiling, and pathway analyses to detect heterogeneity, identify driver mutations, and explore possible mechanisms of tumor evolution. We showed the heterogeneity of the studied samples, characterized the global copy number alteration profiles, and identified genes whose copy number status and expression levels were aberrant. In particular, we identified a recurrent association between two BRAF(V600E) and BRAF(V600K) mutations and changes in DKK1 gene expression levels, which might indicate an association between the BRAF and WNT pathways. These findings show that the integrated approaches used in the present study can robustly address the challenging issue of tumor heterogeneity in high-throughput profiling.

Close proximity to Igh is a contributing factor to AID-mediated translocations.

Class switch recombination (CSR) has the potential to generate genomic instability in B cells as activation-induced cytidine deaminase (AID), which mediates this process, is known to target many sites outside Igh. Nonetheless we do not fully understand what factors influence AID targeting genome-wide. Given that errors in CSR can lead to dangerous, oncogenic chromosomal translocations it is important to identify the elements that determine which genes are at risk of being "hit" and could be involved in aberrant rearrangements. Here we have investigated the influence of nuclear organization in determining "off-target" activity and the choice of fusion partners. Our studies indicate that the vast majority of known AID-mediated Igh translocation partners are found in chromosomal domains that contact this locus during class switching. Further, these interaction domains can be used to identify other genes that are hit by AID.

PCGF homologs, CBX proteins, and RYBP define functionally distinct PRC1 family complexes.

The heterogeneous nature of mammalian PRC1 complexes has hindered our understanding of their biological functions. Here, we present a comprehensive proteomic and genomic analysis that uncovered six major groups of PRC1 complexes, each containing a distinct PCGF subunit, a RING1A/B ubiquitin ligase, and a unique set of associated polypeptides. These PRC1 complexes differ in their genomic localization, and only a small subset colocalize with H3K27me3. Further biochemical dissection revealed that the six PCGF-RING1A/B combinations form multiple complexes through association with RYBP or its homolog YAF2, which prevents the incorporation of other canonical PRC1 subunits, such as CBX, PHC, and SCM. Although both RYBP/YAF2- and CBX/PHC/SCM-containing complexes compact chromatin, only RYBP stimulates the activity of RING1B toward H2AK119ub1, suggesting a central role in PRC1 function. Knockdown of RYBP in embryonic stem cells compromised their ability to form embryoid bodies, likely because of defects in cell proliferation and maintenance of H2AK119ub1 levels.