Erythro- and megakaryoblastic leukemias induced by the murine retrovirus Graffi

Gene expression profiling MOE 430.2

GEO:GSE12581

Additional analyses and Downloadable Excel files


Veronique Voisin, Eric Rassart (Université du Québec à Montréal, Canada)


  1. Introduction: Leukemias induced by the murine retrovirus Graffi (Voisin et al. J.Virol.)
  2. About 50 years ago, the Graffi murine leukemia virus (MuLV) was isolated by Prof. Arnold Graffi at the Cancer Research Institute of Berlin.This retrovirus is able to induce leukemia when inoculated into newborn mice after a latency period. The laboratory of Dr. Eric Rassart cloned 2 non-defective ecotropic retroviral genomes from NIH-3T3 cell lines chronically infected with the original Graffi acellular extract. These 2 viral variants were called GV-1.2 and GV-1.4. Recently, we fully characterized the leukemias induced by GV-1.2 and GV-1.4 in different strains of mice. We found that the Graffi MuLV is a multipotent retrovirus capable of inducing both lymphoid (T- and B-cell) and non-lymphoid (myeloid, erythroid, megakaryoblastic) leukemias. Many of these leukemias are very complex with concomitant expression of different hematopoietic lineages. Interestingly, a high percentage of megakaryoblastic leukemias, a type of leukemia rarely observed with MuLVs, arose in the FVB/n strain of mice. The genetic background of the mice and the viral variant that was inoculated (GV-1.2 or Gv-1.4) influenced the types of leukemias that developed in the mice.

  3. Gene profiling experiments
  4. Gene profiling experiments were performed and gene signatures of the T-cell, B-cell, erythro- and megakaryoblastic leukemias induced by the murine retrovirus Graffi were obtained. NFS mice were inoculated with viral particles of Graffi GV-1.4. Total RNAs of 3 T-cell leukemias and 1 T-cell control (pool of 12 mice),3 B-cell leukemias and 1 B-cell control (pool of 12 mice), 1 myeloid leukemia, 3 erythroid leukemias and 1 erythroid control(pool of 12 mice) and 3 megakaryoblastic leukemias were extracted from purified leukemic populations and were hybridized on Affymetrix Genechips MOE 430.2.The dataset was deposited at GEO (NCBI, accession number GSE12581). Affymetrix MicroArray Suite version 5.0 was used to scan and quantify the arrays.Normalization of gene expression data were performed using the Bioconductor implementation of RMA (Robust Multi Array, B. Bolstad, University of California, Berkeley) available from the Flexarray software (Blazejczyk, M., M. Miron, and R. Nadon. 2007. FlexArray: A statistical data analysis software for gene expression microarrays) and compared to the Affymetrix MAS 5.0 results. For each probeset and for each sample, the deviation to the mean of the fluorescence intensity (RMA results) was calculated. A positive deviation is synonym of an over-expression compared to the mean and a negative deviation is synonym of an under-expression compared to the mean. Unsupervised learning methods (PCA, available form the Flexarray software) were utilized to examine the pattern of the data. The result of the PCA analysis presented here show that leukemias of a same type and related control samples cluster well together:

    This clustering result shows that gene signatures specific of the each type of leukemias can be extracted from the data. Significance Analysis of microarrays (SAM, available from the Flexarray software) was performed using using the normalized data of the 45000 probesets. Data with p-values equal or below 0.01 and False Discover Rates (Benjamini Hochberg) equal or below 0.20 were included in further analyses. The data were ranked in decreasing order of the SAM d-score to obtain the list of the differentially expressed genes.Results of the analyses of the erythroid and megakaryoblastic signatures are summarized in a manuscript in preparation (Erythro- and megakaryoblastic leukemias induced by the murine retrovirus Graffi).The dataset (RMA raw data and .CEL files) was deposited at GEO (NCBI, accession number GSE12581).The complete data of the different erythroid and megakaryoblastic signatures can be downloaded from this website.

  5. Megakaryoblastic specific signatures
    1. Validation of the megakaryoblastic phenotype
    2. The Graffi-virus induced megakaryoblastic leukemias are CD41+Kit+ or CD41-Kit+. The proteins coded by Itga2b (CD41) and Itgb3 (CD61) form a heterodimer called integrin- alpha-IIb-beta-IIIb. This integrin is the fibrinogen receptor that is responsible for platelet aggregation. This aggregation is followed by coagulation and by leukocytes attraction to the site. Platelet aggregation is preceded by platelet activation and granules secretion. The proteins coded by Cd151, Cd9, F11r, Selp, Tln1 are known to interact with the integrin alpha-IIb-beta-IIIb. The proteins coded by Vasp, Syk, Rap1b are regulated and/or regulate the integrin alpha-IIb-beta-IIIb.

      (click twice on the image to enlarge it)

    3. Megakaryoblastic specific signatures: functional classification
    4. In order to functionally characterized the genes of the Mk and MkE1 signatures, the GO terms were associated with the probesets. The probesets were organized into 4 functional class subsets: cycle/cell growth/development/angiogenesis/DNA repair, transcriptional regulation, signalling and aminoacid phosphorylation or dephosphorylation. The next 2 links provide the complete tables (the probesets with GO terms that do not belong to the 4 described functional classes are not included and only one probeset per transcript is represented).

      Mk signature functional classification:

      (click on the image to view the complete table)


      MkMB signature functional classification:

      (click on the image to view the complete table)

    5. Megakaryoblastic specific signature: complete database
    6. This database contains the list of all the genes over-expressed in the megakaryoblastic leukemias.This database can be downloaded and contains the complete data: click on the image, save and open it in Excel. The security level of the Excel macro has to be set to low in order to be able to use the database: in Excel, go to Tools menu, Macro, and set the security level to low. The workbook contains 4 sheets named MK ONLY, MKE1, MKM and MKMB respectively. MK ONLY corresponds to genes over-expressed in the 3 megakaryoblastic leukemias Mk1, Mk2 and Mk3. MKE1 corresponds to genes over-expressed in the 3 megakaryoblastic leukemias and in the erythroleukemia E1. MKM corresponds to genes over-expressed in the 3 megakaryoblastic leukemias and in the myeloid leukemia. MKMB correspond to genes over-expressed in the 3 megakaryoblastic leukemias, the myeloid one and the B-cell leukemias, B1, B2 and B3. The data are organized in descending order: the first probeset has the best SAM result (d-score). The left part of the display window contains the probeset number, gene title and symbol and Unigene ID followed by the SAM results. The right part of the display window shows the ratios of fluorescence intensity calculated from the RMA results for each leukemia. A bar chart illustrates the results. On the left part of the RMA results are the Affymetrix MAS5.0 results (Call signals). It is possible to navigate through the database using the Probeset, Gene Title and Probeset Order dropdown lists or using the scrollbar of the RMA results.

      (click on the image to download the database)

  6. Erythroid specific signatures
    1. Erythroid known genes: heme biosynthesis
    2. The Graffi-virus induced erythroid leukemias are Ter119+(Gypa)CD71+(Tfrc). The red blood cells are constituted of 85% of hemoglobin molecules. Each hemoglobin molecule bears 2 alpha-globin and 2 beta-globin chains. Each globin chain contains a heme linked to an iron molecule that binds oxygen. The heme biosynthesis occurs during the differentiation processus of the erythroid lineage at the erythroblast stage.

      (click twice on the image to enlarge it)

    3. Erythroid specific signature: functional classification
    4. The E and EMk3 signatures have been ordered in 4 functional classes (cycle/cell growth/development/ angiogenesis/DNA repair, transcriptional regulation, signalling and aminoacid phosphorylation or dephosphorylation) as for the megakaryoblastic leukemias (section III.3). Only the probesets that are over-expressed in the erythroid leukemias in comparison to the control sample (columns E1-C, E2-C, E3-C) are represented.

      (click on the image to view the complete table)

    5. Erythroid specific signature: complete database
    6. This database contains two worksheets: one for the genes over-expressed in the 3 erythroid leukemias, E1, E2 and E3 (E Only) and a second one for the genes over-expressed in E1, E2, E3 and in the megakaryoblastic leukemias Mk3 (EMK3). This database can be downloaded and contains the complete data: click on the image, save and open it in Excel. The security level of the Excel macro has to be set to low in order to be able to use the database: in Excel, go to Tools menu, Macro, and set the security level to low.The data are organized in descending order: the first probeset has the best SAM result (d-score). The left part of the display window contains the probeset number, gene title and symbol and Unigene ID followed by the SAM results. The right part of the display window shows the ratios of fluorescence intensity calculated from the RMA results for each leukemia. A bar chart illustrates the results. On the left part of the RMA results are the Affymetrix MAS5.0 results (Call signals). It is possible to navigate through the database using the Probeset, Gene Title and Probeset Order dropdown lists or using the scrollbar of the RMA results.

      (click on the image to download the database)

  7. Erythroid and Megakaryoblastic specific signature: complete database
  8. This database contains list of genes over-expressed both in the erythroid and megakaryoblastic leukemias. This database can be downloaded and contains the complete data: click on the image, save and open it in Excel. The security level of the Excel macro has to be set to low in order to be able to use the database: in Excel, go to Tools menu, Macro, and set the security level to low.The data are organized in descending order: the first probeset has the best SAM result (d-score). The left part of the display window contains the probeset number, gene title and symbol and Unigene ID followed by the SAM results. The right part of the display window shows the ratios of fluorescence intensity calculated from the RMA results for each leukemia. A bar chart illustrates the results. On the left part of the RMA results are the Affymetrix MAS5.0 results (Call signals). It is possible to navigate through the database using the Probeset, Gene Title and Probeset Order dropdown lists or using the scrollbar of the RMA results.

    r
    (click on the image to download the database)

Web site created by Atsu Vovor and Veronique Voisin.
VBA Interface created by Atsu Vovor.

Last modified 06/June/2009