Data supplements

Article Figures & Data


  • Figure 1

    Gene expression analysis. (A) Expression profiling data were converted into a heatmap using CLUSTER and TREEVIEW software (red indicates high; green, low; black, medium expression level). Data show homeobox genes differentially expressed in HL and NHL cell lines. Genes showing significant up-regulation/down-regulation in HL (HOXC6, SHOX2, HOXB8, HOXB9, HOXB4, SATB1, HOP/PAX5, POU2F2), shown in red/green, respectively, were subsequently analyzed by RT-PCR, depicted in panel B together with HLXB9, genes located within the HOXB cluster (HOXB13, PRAC, PRAC2, mir-196a1) and involved in HOXB9 transcription (FGF2). Expression of upstream binding factor (UBF) was used as positive control for cDNA quality. NTC indicates no template control.

  • Figure 2

    RT-PCR analysis of HOXB9 after stimulation of HL cell lines. (A) HDLM-2 and KM-H2 were treated with 200 μM genistein and 14 mM NFκB inhibitor for 12 hours. In contrast to HDLM-2, genistein treatment reduced HOXB9 expression in KM-H2 cells, whereas NFκB inhibitor showed no effect. Antisense oligonucleotide (AS) treatment directed against E2F3A in SUP-HD1 and KM-H2 cells affected HOXB9 reduction relative to control oligonucleotide, indicating an activatory influence for this protein. (B) HDLM-2 and L-428 cells were treated with the following inhibitors of signaling pathways for 12 hours: 1 μM wortmannin (PI3K), 100 μM AG490 (JAK/STAT), 50 μM PD98059, and 100 μM SB202190 (ERK1/2 MAPK). HOXB9 showed reduction in HDLM-2 cells only, following MAPK pathway inhibition, indicating stimulation by this pathway. (C) KM-H2 cells were treated with 10 μg/mL inhibitory antibody directed against FGF2. HDLM-2 and L-540 cells were stimulated with 10 ng/mL recombinant FGF2 protein. HOXB9 expression decreased in KM-H2 and increased in HDLM-2 and L-540 cells. Data indicate an activatory influence of FGF2 on HOXB9 expression. (D) Influence of PKC on HOXB9 expression was analyzed by treating HDLM-2 and KM-H2 cells with a PKC activator TPA (100 nM) or PKC inhibitor calphostin C. Data show the absence of PKC activity in both cell lines. PKC activation reduced HOXB9 expression in KM-H2 but not in HDLM-2. NTC indicates no template control. For cDNA control the expression of ets variant gene 6 (TEL) or UBF was analyzed.

  • Figure 3

    Quantitative RT-PCR analysis after knockdown of BMI1. HDLM-2 and KM-H2 cells were lentivirally transfected with RNAi construct directed against BMI1. RQ-PCR analysis indicated an about 2-fold reduction in BMI1 expression in comparison to GAPDH. Correspondingly, expression of both HOXB9 and HOXB13 increased about 9-fold and 3-fold in HDLM-2 and KM-H2, respectively. Experiments were performed twice with similar results; error bars indicate SD.

  • Figure 4

    Immunostaining of phospho-ERK5 in HL cell lines and lymph nodes. (A) Cytospins of HL cell lines (HDLM-2, L-428, L-540) and of NHL cell line (SC-1) were stained by immunofluorescence with phospho-ERK5 antibody (magnification, × 630). In contrast to SC-1, HL cells show intensive nuclear staining. DAPI counterstaining indicates the nuclei. (B) The HL cell line KM-H2 was treated with 10 μg/mL anti-FGF2 antibody or with 200 μM tyrosine kinase inhibitor genistein and subsequently stained with anti–phospho-ERK5 (magnification, × 630). In contrast to control cells, treated cells show weak staining of nuclei indicating that (1) FGF2 signaling is mediated by ERK5 and (2) genistein inhibits this pathway. (C) Lymph nodes obtained from 3 patients with HL (HL1, HL8, HL10) stained with anti–phospho-ERK5 (magnification, × 630) show positive giant cells, indicating a constitutively active ERK5 pathway in H/RS cells.

  • Figure 5

    Functional analysis of HOXB9 expression in HL cell lines. (A) HL cells (HDLM-2 and KM-H2) and NHL cells (SC-1) were treated with antisense oligonucleotide (AS) directed against HOXB9. In contrast to SC-1, HL cell lines showed reduced growth after 3 days of incubation with AS-HOXB9. (B) KM-H2 cells were electroporated with siRNA directed against HOXB9 or control siRNA and were subsequently incubated in the presence/absence of fetal bovine serum (FBS) or with TPA. Cell counts indicate 60% reduction of viability in the absence of FBS and about 25% in the other settings. Cell counts are related to control siRNA treatment in the presence of FBS. (C) In KM-H2 and HELA cells HOXB9 was overexpressed using expression construct pcHOXB9 in the presence or absence of FBS, respectively. Cell counts of treatments with control vector pcDNA3 in the presence of FBS were set at 100%. In the absence of FBS, both cell lines showed enhanced growth/survival following HOXB9 overexpression. In the presence of FBS only KM-H2 cells showed increased cell counts, indicating cell-type–specific differences. All experiments were performed twice with similar results; error bars indicate SD.


  • Table 1

    Gene expression analysis in cell lines by RT-PCR

    Other BCL
    Solid tumors
    • ALCL indicates anaplastic large cell lymphoma; MLBCL, mediastinal large B-cell lymphoma; DLBCL, diffuse large B-cell lymphoma; FL, follicular lymphoma; MBCL, mediastinal B-cell lymphoma; BL, Burkitt lymphoma; NKTL, natural killer T-cell lymphoma; BCL, B-cell lymphoma/leukemia; TCL, T-cell leukemia; PBC, peripheral blood cells (PBCs); NE, no tested expression; +, positive expression; ++, strong expression; –, negative expression.

    • * HOXB4 expression was judged quantitatively.

    • Lymphoma/leukemia entities with at least 50% positive tested cell lines.

  • Table 2

    ELISA analysis of FGF2 protein

    Cell lineFGF2, ng/mL
    HDLM-211 ± 6
    KM-H2228 ± 99
    L-123612 ± 8
    L-4284 ± 5
    L-54010 ± 13
    SUP-HD14 ± 1
    SC-12 ± 2
    RI-10 ± 0
    RC-K831 ± 2
    • FGF2 protein amounts (± SD) were determined by ELISA within cell-culture medium indicated in 1 × 106 cells for 24 hours.