Molecular screening program to select molecular-based recommended therapies for metastatic cancer patients: analysis from the ProfiLER trial.

BACKGROUND: Antitumor activity of molecular-targeted agents is guided by the presence of documented genomic alteration in specific histological subtypes. We aim to explore the feasibility, efficacy and therapeutic impact of molecular profiling in routine setting. PATIENTS AND METHODS: This multicentric prospective study enrolled adult or pediatric patients with solid or hematological advanced cancer previously treated in advanced/metastatic setting and noneligible to curative treatment. Each molecular profile was established on tumor, relapse or biopsies, and reviewed by a molecular tumor board (MTB) to identify molecular-based recommended therapies (MBRT). The main outcome was to assess the incidence rate of genomic mutations in routine setting, across specific histological types. Secondary objectives included a description of patients with actionable alterations and for whom MBRT was initiated, and overall response rate. RESULTS: Four centers included 2579 patients from February 2013 to February 2017, and the MTB reviewed the molecular profiles achieved for 1980 (76.8%) patients. The most frequently altered genes were CDKN2A (N = 181, 7%), KRAS (N = 177, 7%), PIK3CA (N = 185, 7%), and CCND1 (N = 104, 4%). An MBRT was recommended for 699/2579 patients (27%), and only 163/2579 patients (6%) received at least one MBRT. Out of the 182 lines of MBRT initiated, 23 (13%) partial responses were observed. However, only 0.9% of the whole cohort experienced an objective response. CONCLUSION: An MBRT was provided for 27% of patients in our study, but only 6% of patients actually received matched therapy with an overall response rate of 0.9%. Molecular screening should not be used at present to guide decision-making in routine clinical practice outside of clinical trials.This trial is registered with ClinicalTrials.gov, number NCT01774409.

The immune microenvironment of HPV-negative oral squamous cell carcinoma from never-smokers and never-drinkers patients suggests higher clinical benefit of IDO1 and PD1/PD-L1 blockade.

BACKGROUND: Never-smokers and never-drinkers patients (NSND) suffering from oral squamous cell carcinoma (OSCC) are epidemiologically different from smokers drinkers (SD). We therefore hypothesized that they harbored distinct targetable molecular alterations. PATIENTS AND METHODS: Data from The Cancer Genome Atlas (TCGA) (discovery set), Gene Expression Omnibus and Centre Léon Bérard (CLB) (three validation sets) with available gene expression profiles of HPV-negative OSCC from NSND and SD were mined. Protein expression profiles and genomic alterations were also analyzed from TCGA, and a functional pathway enrichment analysis was carried out. Formalin-fixed paraffin-embedded samples from 44 OSCC including 20 NSND and 24 SD treated at CLB were retrospectively collected to perform targeted-sequencing of 2559 transcripts (HTG EdgeSeq system), and CD3, CD4, CD8, IDO1, and PD-L1 expression analyses by immunohistochemistry (IHC). Enrichment of a six-gene interferon-γ signature of clinical response to pembrozulimab (PD-1 inhibitor) was evaluated in each sample from all cohorts, using the single sample gene set enrichment analysis method. RESULTS: A total of 854 genes and 29 proteins were found to be differentially expressed between NSND and SD in TCGA. Functional pathway analysis highlighted an overall enrichment for immune-related pathways in OSCC from NSND, especially involving T-cell activation. Interferon-γ response and PD1 signaling were strongly enriched in NSND. IDO1 and PD-L1 were overexpressed and the score of response to pembrolizumab was higher in NSND than in SD, although the mutational load was lower in NSND. IHC analyses in the CLB cohort evidenced IDO1 and PD-L1 overexpression in tumor cells that was associated with a higher rate of tumor-infiltrating T-cells in NSND compared with SD. CONCLUSION: The main biological and actionable difference between OSCC from NSND and SD lies in the immune microenvironment, suggesting a higher clinical benefit of PD-L1 and IDO1 inhibition in OSCC from NSND.

EZH2 alterations in follicular lymphoma: biological and clinical correlations.

The histone methyltransferase EZH2 has an essential role in the development of follicular lymphoma (FL). Recurrent gain-of-function mutations in EZH2 have been described in 25% of FL patients and induce aberrant methylation of histone H3 lysine 27 (H3K27). We evaluated the role of EZH2 genomic gains in FL biology. Using RNA sequencing, Sanger sequencing and SNP-arrays, the mutation status, copy-number and gene-expression profiles of EZH2 were assessed in a cohort of 159 FL patients from the PRIMA trial. Immunohistochemical (IHC) EZH2 expression (n=55) and H3K27 methylation (n=63) profiles were also evaluated. In total, 37% of patients (59/159) harbored an alteration in the EZH2 gene (mutation n=46, gain n=23). Both types of alterations were associated with highly similar transcriptional changes, with increased proliferation programs. An H3K27me3/me2 IHC score fully distinguished mutated from wild-type samples, showing its applicability as surrogate for EZH2 mutation analysis. However, this score did not predict the presence of gains at the EZH2 locus. The presence of an EZH2 genetic alteration was an independent factor associated with a longer progression-free survival (hazard ratio 0.58, 95% confidence interval 0.36-0.93, P=0.025). We propose that the copy-number status of EZH2 should also be considered when evaluating patient stratification and selecting patients for EZH2 inhibitor-targeted therapies.

The complete genome sequence of the murine respiratory pathogen Mycoplasma pulmonis.

Mycoplasma pulmonis is a wall-less eubacterium belonging to the Mollicutes (trivial name, mycoplasmas) and responsible for murine respiratory diseases. The genome of strain UAB CTIP is composed of a single circular 963 879 bp chromosome with a G + C content of 26.6 mol%, i.e. the lowest reported among bacteria, Ureaplasma urealyticum apart. This genome contains 782 putative coding sequences (CDSs) covering 91.4% of its length and a function could be assigned to 486 CDSs whilst 92 matched the gene sequences of hypothetical proteins, leaving 204 CDSs without significant database match. The genome contains a single set of rRNA genes and only 29 tRNAs genes. The replication origin oriC was localized by sequence analysis and by using the G + C skew method. Sequence polymorphisms within stretches of repeated nucleotides generate phase-variable protein antigens whilst a recombinase gene is likely to catalyse the site-specific DNA inversions in major M.pulmonis surface antigens. Furthermore, a hemolysin, secreted nucleases and a glyco-protease are predicted virulence factors. Surprisingly, several of the genes previously reported to be essential for a self-replicating minimal cell are missing in the M.pulmonis genome although this one is larger than the other mycoplasma genomes fully sequenced until now.

Evolutionary role of restriction/modification systems as revealed by comparative genome analysis.

Type II restriction modification systems (RMSs) have been regarded either as defense tools or as molecular parasites of bacteria. We extensively analyzed their evolutionary role from the study of their impact in the complete genomes of 26 bacteria and 35 phages in terms of palindrome avoidance. This analysis reveals that palindrome avoidance is not universally spread among bacterial species and that it does not correlate with taxonomic proximity. Palindrome avoidance is also not universal among bacteriophage, even when their hosts code for RMSs, and depends strongly on the genetic material of the phage. Interestingly, palindrome avoidance is intimately correlated with the infective behavior of the phage. We observe that the degree of palindrome and restriction site avoidance is significantly and consistently less important in phages than in their bacterial hosts. This result brings to the fore a larger selective load for palindrome and restriction site avoidance on the bacterial hosts than on their infecting phages. It is then consistent with a view where type II RMSs are considered as parasites possibly at the verge of mutualism. As a consequence, RMSs constitute a nontrivial third player in the host-parasite relationship between bacteria and phages.

Searching genomes for sequences with the potential to form intrastrand triple helices.

The canonical double-helix form of DNA is thought to predominate both in dilute solution and in living cells. Sequence-dependent fluctuations in local DNA shape occur within the double helix. Besides these relatively modest variations in shape, more extreme and remarkable structures have been detected in which some bases become unpaired. Examples include unusual three-stranded structures such as H-DNA. Certain RNA and DNA strands can also fold onto themselves to form intrastrand triplexes. Although they have been extensively studied in vitro, it remains unknown whether nucleic acid triplexes play natural roles in cells. If natural nucleic acid triplexes were identified in cells, much could be learned by examining the formation, stabilization, and function of such structures. With these goals in mind, we adapted a pattern-recognition program to search genetic databases for a type of potential triplex structure whose presence in genomes has not been previously investigated. We term these sequences Potential Intrastrand Triplex (PIT) elements. The formation of an intrastrand triplex requires three consecutive sequence domains with appropriate symmetry along a single nucleic acid strand. It is remarkable that we discovered multiple copies of sequence elements with the potential to form one particular class of intrastrand triplexes in the fully sequenced genomes of several bacteria. We then focused on the characterization of the 25 copies of a particular approximately 37 nt PIT sequence detected in Escherichia coli. Through biochemical studies, we demonstrate that an isolated DNA strand from this family of E. coli PIT elements forms a stable intrastrand triplex at physiological temperature and pH in the presence of physiological concentrations of Mg(2+).

Analysis of intrachromosomal duplications in yeast Saccharomyces cerevisiae: a possible model for their origin.

The complete genome of the yeast Saccharomyces cerevisiae was investigated for intrachromosomal duplications at the level of nucleotide sequences. The analysis was performed by looking for long approximate repeats (from 30 to 3,885 bp) present on each of the chromosomes. We show that direct and inverted repeats exhibit very different characteristics: the two copies of direct repeats are more similar and longer than those of inverted repeats. Furthermore, contrary to the inverted repeats, a large majority of direct repeats appear to be closely spaced. The distance (delta) between the two copies is generally smaller than 1 kb. Further analysis of these "close direct repeats" shows a negative correlation between delta and the percentage of identity between the two copies, and a positive correlation between delta and repeat length. Moreover, contrary to the other categories of repeats, close direct repeats are mostly located within coding sequences (CDSs). We propose two hypotheses in order to interpret these observations: first, the deletion/conversion rate is negatively correlated with delta; second, there exists an active duplication mechanism which continuously creates close direct repeats, the other intrachromosomal repeats being the result, by chromosomal rearrangements of these "primary repeats."

Implication of gene distribution in the bacterial chromosome for the bacterial cell factory.

As bacterial genome sequences accumulate, more and more pieces of data suggest that there is a significant correlation between the distribution of genes along the chromosome and the physical architecture of the cell, suggesting that the map of the cell is in the chromosome. Considering sequences and experimental data indicative of cell compartmentalisation, mRNA folding and turnover, as well as known structural features of protein and membrane complexes, we show that preliminary in silico analysis of whole genome sequences strongly substantiates this hypothesis. If there is a correlation between the genome sequence and the cell architecture, it must derive from some selection pressure in the organisms growing in the wild. As a consequence, the underlying constraints should be optimised in genetically modified organisms if one is to expect high product yields. Consequences in terms of gene expression for biotechnology are straightforward: knocking genes out and in genomes should not be randomly performed, but should follow the rules of chromosome organisation.

Detecting and analyzing DNA sequencing errors: toward a higher quality of the Bacillus subtilis genome sequence.

During the determination of a DNA sequence, the introduction of artifactual frameshifts and/or in-frame stop codons in putative genes can lead to misprediction of gene products. Detection of such errors with a method based on protein similarity matching is only possible when related sequences are available in databases. Here, we present a method to detect frameshift errors in DNA sequences that is based on the intrinsic properties of the coding sequences. It combines the results of two analyses, the search for translational initiation/termination sites and the prediction of coding regions. This method was used to screen the complete Bacillus subtilis genome sequence and the regions flanking putative errors were resequenced for verification. This procedure allowed us to correct the sequence and to analyze in detail the nature of the errors. Interestingly, in several cases in-frame termination codons or frameshifts were not sequencing errors but confirmed to be present in the chromosome, indicating that the genes are either nonfunctional (pseudogenes) or subject to regulatory processes such as programmed translational frameshifts. The method can be used for checking the quality of the sequences produced by any prokaryotic genome sequencing project.

Analysis of long repeats in bacterial genomes reveals alternative evolutionary mechanisms in Bacillus subtilis and other competent prokaryotes.

Prokaryotic genomes seem to be optimized toward compactness and have therefore been thought to lack long redundant DNA sequences. However, we identified a large number of long strict repeats in eight prokaryotic complete genomes and found that their density is negatively correlated with genome size. A detailed analysis of the long repeats present in the genome of Bacillus subtilis revealed a very strict constraint on the spatial distribution of repeats in this genome. We interpret this as the hallmark of selection processes leading to the addition of new genetic information. Such addition is independent of insertion sequences and relies on the nonspecific DNA uptake by the competent cell and its subsequent integration in the chromosome in a circular form through a Campbell-like mechanism. Similar patterns are found in other competent genomes of Gram-negative bacteria and Archaea, suggesting a similar evolutionary mechanism. The correlation of the spatial distribution of repeats and the absence of insertion sequences in a genome may indicate, in the framework of our model, that mechanisms aiming at their avoidance/elimination have been developed.

Translation in Bacillus subtilis: roles and trends of initiation and termination, insights from a genome analysis.

We analysed the Bacillus subtilis protein coding sequences termini, and compared it to other genomes. The analysis focused on signals, com-positional biases of nucleotides, oligonucleotides, codons and amino acids and mRNA secondary structure. AUG is the preferred start codon in all genomes, independent of their G+C content, and seems to induce less stable mRNA structures. However, it is not conserved between homologous genes neither is it preferred in highly expressed genes. In B.subtilis the ribosome binding site is very strong. We found that downstream boxes do not seem to exist either in Escherichia coli or in B.subtilis. UAA stop codon usage is correlated with the G+C content and is strongly selected in highly expressed genes. We found less stable mRNA structures at both termini, which we related to mRNA-ribosome and mRNA-release-factor interactions. This pattern seems to impose a peculiar A-rich nucleotide and codon usage bias in these regions. Finally the analysis of all proteins from B.subtilis revealed a similar amino acid bias near both termini of proteins consisting of over-representation of hydrophilic residues. This bias near the stop codon is partially release-factor specific.

Universal replication biases in bacteria.

Analysis of 15 complete bacterial chromosomes revealed important biases in gene organization. Strong compositional asymmetries between the genes lying on the leading versus lagging strands were observed at the level of nucleotides, codons and, surprisingly, amino acids. For some species, the bias is so high that the sole knowledge of a protein sequence allows one to predict with almost no errors whether the gene is transcribed from one strand or the other. Furthermore, we show that these biases are not species specific but appear to be universal. These findings may have important consequences in our understanding of fundamental biological processes in bacteria, such as replication fidelity, codon usage in genes and even amino acid usage in proteins.

Imagene: an integrated computer environment for sequence annotation and analysis.

MOTIVATION: To be fully and efficiently exploited, data coming from sequencing projects together with specific sequence analysis tools need to be integrated within reliable data management systems. Systems designed to manage genome data and analysis tend to give a greater importance either to the data storage or to the methodological aspect, but lack a complete integration of both components. RESULTS: This paper presents a co-operative computer environment (called Imagenetrade mark) dedicated to genomic sequence analysis and annotation. Imagene has been developed by using an object-based model. Thanks to this representation, the user can directly manipulate familiar data objects through icons or lists. Imagene also incorporates a solving engine in order to manage analysis tasks. A global task is solved by successive divisions into smaller sub-tasks. During program execution, these sub-tasks are graphically displayed to the user and may be further re-started at any point after task completion. In this sense, Imagene is more transparent to the user than a traditional menu-driven package. Imagene also provides a user interface to display, on the same screen, the results produced by several tasks, together with the capability to annotate these results easily. In its current form, Imagene has been designed particularly for use in microbial sequencing projects. AVAILABILITY: Imagene best runs on SGI (Irix 6.3 or higher) workstations. It is distributed free of charge on a CD-ROM, but requires some Ilog licensed software to run. Some modules also require separate license agreements. Please contact the authors for specific academic conditions and other Unix platforms. CONTACT: imagene home page: http://wwwabi.snv.jussieu.fr/imagene

Functional and evolutionary roles of long repeats in prokaryotes.

Most recently published complete bacterial genomes have revealed unexpectedly high numbers of long strict repeats. In this article we discuss the various functional and evolutionary roles of these repeats, focusing in particular on their role in terms of genome stability, gene transfer, and antigenic variation.

Three distinct sub-nuclear populations of HMG-I protein of different properties revealed by co-localization image analysis.

We have studied the nuclear distribution of the non-histone HMG-I protein by indirect immunofluorescence in several human and murine somatic cell lines and in growing mouse oocytes. We show that HMG-I, a high mobility-group protein which interacts in vitro with the minor groove of AT-rich B-DNA, is found exclusively in the nucleus and that this localization corresponds to a complex distribution. By comparing the HMG-I-dependent fluorescence signal with the chromatin density determined by Hoechst 33342 or propidium iodide staining, we present evidence for the existence of three HMG-I sub-populations whose contribution to the total fluorescence can be determined using a newly developed quantitative co-localization image analysis program: foci that correspond to regions of heterochromatin, intense dots located within decondensed chromatin, and a more diffuse component extending throughout the nucleoplasm. In addition, we show that these sub-populations differ in their sensitivity to nuclease digestion and in vivo displacement by the minor-groove binder Hoechst 33342. Finally, double immunolabeling of RNA polymerase II-dependent transcription and HMG-I shows that the intense dots are not correlated with sites of high transcriptional activity. We discuss the possibility that these three sub-populations reflect distinct and separable biological functions of the HMG-I protein.

Oligonucleotide bias in Bacillus subtilis: general trends and taxonomic comparisons.

We present a general analysis of oligonucleotide usage in the complete genome of Bacillus subtilis . Several datasets were built in order to assign various biological contexts to the biased use of words and to reveal local asymmetries in word usage that may be coupled with replication, the control of gene expression and the restriction/modification system. This analysis was complemented by cross-comparisons with the complete genomes of Escherichia coli , Haemophilus influenzae and Methanococcus jannaschii . We have observed a large number of biased oligonucleotides for words of size up to 8, throughout the datasets and species, indicating that such long strict words play an important role as biological signals. We speculate that some of them are involved in interactions with DNA and/or RNA polymerases. An extensive analysis of palindrome abundances and distributions provides the surprising result that prophage-like elements embedded in the genome exhibit a smaller avoidance of restriction sites. This may reinforce a recently proposed hypothesis of a selfish gene phenomena in the transfer of restriction/modification systems in bacteria.

Pairwise and multiple identification of three-dimensional common substructures in proteins.

In this paper, we present an algorithm to find three-dimensional substructures common to two or more molecules. The basic algorithm is devoted to pairwise structural comparison. Given two sets of atomic coordinates, it finds the largest subsets of atoms which are "similar" in the sense that all internal distances are approximately conserved. The basic idea of the algorithm is to recursively build subsets of increasing sizes, combining two sets of size k to build a set of size k + 1. The algorithm can be used "as is" for small molecules or local parts of proteins (about 30 atoms). When a high number of atoms is involved, we use a two step procedure. First we look for common "local" fragments by using the previous algorithm, and then we gather these fragments by using a Branch and Bound technique. We also extend the basic algorithm to perform multiple comparisons, by using one of the structures as a reference point (pivot) to which all other structures are compared. The solution is the largest subsets of atoms common to the pivot and at least q other structures. Although both algorithms are theoretically exponential in the number of atoms, experiments performed on biological data and using realistic parameters show that the solution is obtained within a few minutes. Finally, an application to the determination of the structural core of seven globins is presented.

Automated multiple analysis of protein structures: application to homology modeling of cytochromes P450.

A computational strategy for homology modeling, using several protein structures comparison, is described. This strategy implies a formalized definition of structural blocks common to several protein structures, a new program to compare these structures simultaneously, and the use of consensus matrices to improve sequence alignment between the structurally known and target proteins. Applying this method to cytochromes P450 led to the definition of 15 substructures common to P450cam, P450BM3, and P450terp, and to proposing a 3D model of P450eryF.