Double-negative (CD27-IgD-) B cells are expanded in NSCLC and inversely correlate with affinity-matured B cell populations.

BACKGROUND: The presence of B cells in early stage non-small cell lung cancer (NSCLC) is associated with longer survival, however, the role these cells play in the generation and maintenance of anti-tumor immunity is unclear. B cells differentiate into a variety of subsets with differing characteristics and functions. To date, there is limited information on the specific B cell subsets found within NSCLC. To better understand the composition of the B cell populations found in NSCLC we have begun characterizing B cells in lung tumors and have detected a population of B cells that are CD79A+CD27-IgD-. These CD27-IgD- (double-negative) B cells have previously been characterized as unconventional memory B cells and have been detected in some autoimmune diseases and in the elderly population but have not been detected previously in tumor tissue. METHODS: A total of 15 fresh untreated NSCLC tumors and 15 matched adjacent lung control tissues were dissociated and analyzed by intracellular flow cytometry to detect the B cell-related markers CD79A, CD27 and IgD. All CD79A+ B cells subsets were classified as either naïve (CD27-IgD+), affinity-matured (CD27+IgD-), early memory/germinal center cells (CD27+IgD+) or double-negative B cells (CD27-IgD-). Association of double-negative B cells with clinical data including gender, age, smoking status, tumor diagnosis and pathologic differentiation status were also examined using the logistic regression analysis for age and student's t-test for all other variables. Associations with other B cell subpopulations were examined using Spearman's rank correlation. RESULTS: We observed that double-negative B cells were frequently abundant in lung tumors compared to normal adjacent controls (13 out of 15 cases), and in some cases made up a substantial proportion of the total B cell compartment. The presence of double-negative cells was also found to be inversely related to the presence of affinity-matured B cells within the tumor, Spearman's coefficient of - 0.76. CONCLUSIONS: This study is the first to observe the presence of CD27-IgD- double-negative B cells in human NSCLC and that this population is inversely correlated with traditional affinity-matured B cell populations.

Alisertib added to rituximab and vincristine is synthetic lethal and potentially curative in mice with aggressive DLBCL co-overexpressing MYC and BCL2.

Pearson correlation coefficient for expression analysis of the Lymphoma/Leukemia Molecular Profiling Project (LLMPP) demonstrated Aurora A and B are highly correlated with MYC in DLBCL and mantle cell lymphoma (MCL), while both Auroras correlate with BCL2 only in DLBCL. Auroras are up-regulated by MYC dysregulation with associated aneuploidy and resistance to microtubule targeted agents such as vincristine. Myc and Bcl2 are differentially expressed in U-2932, TMD-8, OCI-Ly10 and Granta-519, but only U-2932 cells over-express mutated p53. Alisertib [MLN8237 or M], a highly selective small molecule inhibitor of Aurora A kinase, was synergistic with vincristine [VCR] and rituximab [R] for inhibition of cell proliferation, abrogation of cell cycle checkpoints and enhanced apoptosis versus single agent or doublet therapy. A DLBCL (U-2932) mouse model showed tumor growth inhibition (TGI) of ∼ 10-20% (p = 0.001) for M, VCR and M-VCR respectively, while R alone showed ∼ 50% TGI (p = 0.001). M-R and VCR-R led to tumor regression [TR], but relapsed 10 days after discontinuing therapy. In contrast, M-VCR-R demonstrated TR with no relapse >40 days after stopping therapy with a Kaplan-Meier survival of 100%. Genes that are modulated by M-VCR-R (CENP-C, Auroras) play a role in centromere-kinetochore function in an attempt to maintain mitosis in the presence of synthetic lethality. Together, our data suggest that the interaction between alisertib plus VCR plus rituximab is synergistic and synthetic lethal in Myc and Bcl-2 co-expressing DLBCL. Alisertib plus vincristine plus rituximab [M-VCR-R] may represent a new strategy for DLBCL therapy.

Using logistic regression to improve the prognostic value of microarray gene expression data sets: application to early-stage squamous cell carcinoma of the lung and triple negative breast carcinoma.

BACKGROUND: Numerous microarray-based prognostic gene expression signatures of primary neoplasms have been published but often with little concurrence between studies, thus limiting their clinical utility. We describe a methodology using logistic regression, which circumvents limitations of conventional Kaplan Meier analysis. We applied this approach to a thrice-analyzed and published squamous cell carcinoma (SQCC) of the lung data set, with the objective of identifying gene expressions predictive of early death versus long survival in early-stage disease. A similar analysis was applied to a data set of triple negative breast carcinoma cases, which present similar clinical challenges. METHODS: Important to our approach is the selection of homogenous patient groups for comparison. In the lung study, we selected two groups (including only stages I and II), equal in size, of earliest deaths and longest survivors. Genes varying at least four-fold were tested by logistic regression for accuracy of prediction (area under a ROC plot). The gene list was refined by applying two sliding-window analyses and by validations using a leave-one-out approach and model building with validation subsets. In the breast study, a similar logistic regression analysis was used after selecting appropriate cases for comparison. RESULTS: A total of 8594 variable genes were tested for accuracy in predicting earliest deaths versus longest survivors in SQCC. After applying the two sliding window and the leave-one-out analyses, 24 prognostic genes were identified; most of them were B-cell related. When the same data set of stage I and II cases was analyzed using a conventional Kaplan Meier (KM) approach, we identified fewer immune-related genes among the most statistically significant hits; when stage III cases were included, most of the prognostic genes were missed. Interestingly, logistic regression analysis of the breast cancer data set identified many immune-related genes predictive of clinical outcome. CONCLUSIONS: Stratification of cases based on clinical data, careful selection of two groups for comparison, and the application of logistic regression analysis substantially improved predictive accuracy in comparison to conventional KM approaches. B cell-related genes dominated the list of prognostic genes in early stage SQCC of the lung and triple negative breast cancer.

In Vitro Assessment of the Inflammatory Breast Cancer Cell Line SUM 149: Discovery of 2 Single Nucleotide Polymorphisms in the RNase L Gene.

BACKGROUND: Inflammatory breast cancer (IBC) is a rare, highly aggressive form of breast cancer. The mechanism of IBC carcinogenesis remains unknown. We sought to evaluate potential genetic risk factors for IBC and whether or not the IBC cell lines SUM149 and SUM190 demonstrated evidence of viral infection. METHODS: We performed single nucleotide polymorphism (SNP) genotyping for 2 variants of the ribonuclease (RNase) L gene that have been correlated with the risk of prostate cancer due to a possible viral etiology. We evaluated dose-response to treatment with interferon-alpha (IFN-α); and assayed for evidence of the putative human mammary tumor virus (HMTV, which has been implicated in IBC) in SUM149 cells. A bioinformatic analysis was performed to evaluate expression of RNase L in IBC and non-IBC. RESULTS: 2 of 2 IBC cell lines were homozygous for RNase L common missense variants 462 and 541; whereas 2 of 10 non-IBC cell lines were homozygous positive for the 462 variant (p= 0.09) and 0 of 10 non-IBC cell lines were homozygous positive for the 541 variant (p = 0.015). Our real-time polymerase chain reaction (RT-PCR) and Southern blot analysis for sequences of HMTV revealed no evidence of the putative viral genome. CONCLUSION: We discovered 2 SNPs in the RNase L gene that were homozygously present in IBC cell lines. The 462 variant was absent in non-IBC lines. Our discovery of these SNPs present in IBC cell lines suggests a possible biomarker for risk of IBC. We found no evidence of HMTV in SUM149 cells. A query of a panel of human IBC and non-IBC samples showed no difference in RNase L expression. Further studies of the RNase L 462 and 541 variants in IBC tissues are warranted to validate our in vitro findings.

Comparing programs and methods to use for global multiple sequence alignment.

It is difficult to find a global optimal alignment of more than two sequences (and, especially, more than three) that includes matches, mismatches, and gaps and that takes into account the degree of variation in all of the sequences at the same time. Thus, approximate methods are used, such as progressive global alignment, iterative global alignment, alignments based on locally conserved patterns found in the same order in the sequences, statistical methods that generate probabilistic models of the sequences, and multiple sequence alignments produced by graph-based methods. When 10 or more sequences are being compared, it is common to begin by determining sequence similarities between all pairs of sequences in the set. A variety of methods are then available to cluster the sequences into the most related groups or into a phylogenetic tree. This article discusses several of these methods and provides data that compare their utility under various conditions.

Using hidden Markov models to align multiple sequences.

A hidden Markov model (HMM) is a probabilistic model of a multiple sequence alignment (msa) of proteins. In the model, each column of symbols in the alignment is represented by a frequency distribution of the symbols (called a "state"), and insertions and deletions are represented by other states. One moves through the model along a particular path from state to state in a Markov chain (i.e., random choice of next move), trying to match a given sequence. The next matching symbol is chosen from each state, recording its probability (frequency) and also the probability of going to that state from a previous one (the transition probability). State and transition probabilities are multiplied to obtain a probability of the given sequence. The hidden nature of the HMM is due to the lack of information about the value of a specific state, which is instead represented by a probability distribution over all possible values. This article discusses the advantages and disadvantages of HMMs in msa and presents algorithms for calculating an HMM and the conditions for producing the best HMM.

Using progressive methods for global multiple sequence alignment.

Finding a global optimal alignment of more than two sequences that includes matches, mismatches, and gaps and that takes into account the degree of variation in all of the sequences at the same time is especially difficult. The dynamic programming algorithm used for optimal alignment of pairs of sequences can be extended to global alignment of three sequences, but for more than three sequences, only a small number of relatively short sequences may be analyzed. Thus, approximate methods are used for global sequence alignment. One class of these methods is progressive global alignment, which starts with an alignment of the most alike sequences and then builds an alignment by adding more sequences. This article introduces three programs that use progressive alignment methodology.

Using iterative methods for global multiple sequence alignment.

Finding a global optimal alignment of more than two sequences that includes matches, mismatches, and gaps and that takes into account the degree of variation in all of the sequences at the same time is especially difficult. The dynamic programming algorithm used for optimal alignment of pairs of sequences can be extended to global alignment of three sequences, but for more than three sequences, only a small number of relatively short sequences may be analyzed. Thus, approximate methods are used for global alignment. One class of these is iterative global alignment, which makes an initial global alignment of groups of sequences and then revises the alignment to achieve a more reasonable result. This article discusses several iterative alignment methods. In particular, steps are provided for using the Sequence Alignment by Genetic Algorithm (SAGA).

Using multiple sequence alignment editors and formatters.

Sequence alignment editors enable the user to manually edit a multiple sequence alignment (msa) in order to obtain a more reasonable or expected alignment. Editors allow sequences to be reordered and/or modified using the computer's cut and paste commands. They are designed to accept various msa formats and to provide the output file in a suitable user-designated format. Sequence formatters provide various output formatting options, such as color and shading schemes to enhance visualization of residue alignments. The formatters can output files in Postscript, EPS, RTF, and other widely recognized formats, while accepting the standard input formats, such as MSF, ALN, and FASTA. This article introduces a number of sequence alignment editors and formatters, and provides links to sites where they can be found.

Choosing a method for phylogenetic prediction.

INTRODUCTIONThree methods--maximum parsimony, distance, and maximum likelihood--are generally used to find the evolutionary tree or trees that best account for the observed variation in a group of sequences. Each of these methods uses a different type of analysis. Programs based on distance methods are commonly used in the molecular biology laboratory because they are straightforward and can be used with a large number of sequences. Maximum likelihood methods are more challenging and require a greater understanding of the evolutionary models on which they are based. Because they involve so many computational steps and because the number of steps increases dramatically with the number of sequences, maximum likelihood programs are limited to a smaller number of sequences. They can be implemented on a supercomputer in order to analyze a greater number of sequences. This article presents an overview for the researcher who has a set of related sequences and wants to analyze them to predict the best trees that depict the phylogenetic relationships among the sequences.

Maximum parsimony method for phylogenetic prediction.

INTRODUCTIONMaximum parsimony predicts the evolutionary tree or trees that minimize the number of steps required to generate the observed variation in the sequences from common ancestral sequences. For this reason, the method is also sometimes referred to as the minimum evolution method. A multiple sequence alignment (msa) is required to predict which sequence positions are likely to correspond. These positions will appear in vertical columns in the msa. For each aligned position, phylogenetic trees that require the smallest number of evolutionary changes to produce the observed sequence changes from ancestral sequences are identified. This analysis is continued for every position in the sequence alignment. Finally, those trees that produce the smallest number of changes overall for all sequence positions are identified. This method is best suited for sequences that are quite similar and is limited to small numbers of sequences.

Distance methods for phylogenetic prediction.

INTRODUCTIONPhylogenetic analysis of a multiple sequence alignment (msa) can be performed using distance methods, which are based on genetic distances between sequence pairs in an msa. The genetic distance between two sequences is the fraction of aligned positions in which the sequence has been changed. In contrast, sequence identity is the fraction of the aligned positions that are identical. Gaps may be ignored in distance calculations or treated like substitutions. A scoring or substitution matrix may also be used, making the calculation slightly more complicated, although the principle is the same. Sequence pairs that have the smallest distances are "neighbors." On a tree, these sequences share a node or common ancestor position and are each joined to that node by a branch.

The maximum likelihood approach for phylogenetic prediction.

INTRODUCTIONMaximum likelihood (ML) methods are especially useful for phylogenetic prediction when there is considerable variation among the sequences in the multiple sequence alignment (msa) to be analyzed. ML methods start with a simple model, in this case a model of rates of evolutionary change in nucleic acid or protein sequences and tree models that represent a pattern of evolutionary change, and then adjust the model until there is a best fit to the observed data. For phylogenetic analysis, the observed data are the observed sequence variations found within the columns of an msa. The ML method is similar to the maximum parsimony method in that the analysis is performed on each column of an msa.

A test of the markov model of evolution in proteins.

INTRODUCTIONThe percent accepted mutation (PAM) scoring matrix is based on the Dayhoff model of protein evolution, which is a Markov process. In the Markov model of amino acid change, the probability of mutation at each site is independent of the previous history of mutations. Use of this model makes it possible to extrapolate amino acid substitutions observed over a relatively short period of evolutionary time to longer periods of evolutionary time. One criticism of the PAM scoring matrix is that the frequency of amino acid changes that require two nucleotide changes is higher than would be expected by chance. This article describes a test of the Markov model of protein evolution, which shows that the model can be valid if certain changes are made in the way that PAM matrices are calculated.

Comparison of the PAM and BLOSUM Amino Acid Substitution Matrices.

INTRODUCTIONThe choice of a scoring system including scores for matches, mismatches, substitutions, insertions, and deletions influences the alignment of both DNA and protein sequences. To score matches and mismatches in alignments of proteins, it is necessary to know how often one amino acid is substituted for another in related proteins. Percent accepted mutation (PAM) matrices list the likelihood of change from one amino acid to another in homologous protein sequences during evolution and thus are focused on tracking the evolutionary origins of proteins. In contrast, the blocks amino acid substitution matrices (BLOSUM) are based on scoring substitutions found over a range of evolutionary periods. There are important differences in the ways that the PAM and BLOSUM scoring matrices were derived. These differences, which are discussed in this article, should be appreciated when interpreting the results of protein sequence alignments obtained with these matrices.

Studies of varying alignment algorithm, amino Acid scoring matrix, and gap penalties.

INTRODUCTIONComparing different amino acid scoring matrix-gap penalty combinations poses several problems. For example, the analysis often overlooks the purposes of different matrices; e.g., protein family or domain searching, evolutionary analysis, or structural alignment. In the past, gap penalties were usually not published or well known, thus throwing a level of uncertainty into the results. More recently, when investigators publish a new scoring matrix, they usually provide suitable choices for gap penalties that may be used for comparisons with other matrices. This article summarizes a number of reports that have examined combinations of alignment algorithm, scoring matrix, and gap penalties used to align sequences for various purposes.

Using BLOSUM in Sequence Alignments.

INTRODUCTIONThe original Dayhoff percent accepted mutation (PAM) matrices were developed based on a small number of protein sequences and an evolutionary model of protein change. By extrapolating from the observed changes at small evolutionary distances to large ones, it was possible to establish a PAM250 scoring matrix for sequences that were highly divergent. Another approach to finding a scoring matrix for divergent sequences is to start with a more divergent set of sequences and produce a scoring matrix from the substitutions found in those less-related sequences. The blocks amino acid substitution matrices (BLOSUM) scoring matrices were prepared this way. This article explains how BLOSUM scoring matrices were created and how they can best be used.

Using PAM Matrices in Sequence Alignments.

INTRODUCTIONCertain amino acid substitutions commonly occur in related proteins from different species. Because a protein still functions with these substitutions, the substituted amino acids are compatible with protein structure and function. Knowing the types of changes that are most and least common in a large number of proteins can assist with predicting alignments for any set of protein sequences. If related protein sequences are quite similar, they are easy to align, and one can readily determine the single-step amino acid changes. If ancestor relationships among a group of proteins are assessed, the most likely amino acid changes that occurred during evolution can be predicted. This type of analysis was pioneered by Margaret Dayhoff and used by her to produce a type of scoring matrix called a percent accepted mutation (PAM) matrix. This article introduces Dayhoff PAM matrices, explains how they are constructed and how they can be used for sequence alignments, and highlights their strengths and limitations.

Using gaps and gap penalties to optimize pairwise sequence alignments.

INTRODUCTIONTo obtain the best possible alignment between two sequences, it is necessary to include gaps in sequence alignments and use gap penalties. For aligning DNA sequences, a simple positive score for matches and a negative score for mismatches and gaps are most often used. To score matches and mismatches in alignments of proteins, it is necessary to know how often one amino acid is substituted for another in related proteins. In addition, a method is needed to account for insertions and deletions that sometimes appear in related DNA or protein sequences. To accommodate such sequence variations, gaps that appear in sequence alignments are given a negative penalty score reflecting the fact that they are not expected to occur very often. Mathematically speaking, it is very difficult to produce the best-possible alignment, either global or local, unless gaps are included in the alignment. This article discusses how to use gaps and gap penalties to optimize pairwise sequence alignments.

Dot matrix pairwise sequence comparison.

INTRODUCTIONA dot matrix analysis is primarily a method for comparing two sequences to look for possible alignment of characters between the sequences. The method is also used for finding direct or inverted repeats in protein and DNA sequences, and for predicting regions in RNA that are self-complementary and that, therefore, have the potential of forming secondary structure through base-pairing.