Constitutively overexpressed 21 kDa protein in Hodgkin lymphoma and aggressive non-Hodgkin lymphomas identified as cytochrome B5b (CYB5B)

Abstract

Background: We have previously reported a novel constitutively overexpressed 21 kDa protein in Hodgkin Lymphoma (HL) and aggressive Non-Hodgkin Lymphomas (NHL). The objective of the current study was to 1) identify this protein using two independent methods, 2) study the expression of the protein and its encoding mRNA in reactive lymph nodes, normal lymphocytes and CD34+ bone marrow precursor cells, 3) analyse patterns of expression of the protein in tissue microarrays assembled from a large number of diagnostic clinical biopsies from patients with HL, and 4) determine the copy number variation and mutation status of the encoding gene in HL cell lines.

Results: Peptide sequencing by LC-MS/MS and protein identification by protein array screening identified a single protein, CYB5B. No mutations were detected in the CYB5B gene in HL cell lines. Quantitative PCR showed CYB5B gene expression was increased in HL and NHL cell lines. Array CGH using a submegabase resolution tiling array revealed gains in the CYB5B locus in HL cell lines KMH2 and L428. Membrane expression was seen in Reed-Sternberg cells in clinical biopsies from patients with HL but not in reactive lymph nodes. Bone marrow CD34+ precursor cells were CYB5B negative on the cell surface. RT-PCR assays of RNA extracted from T and B cell enriched fractions obtained from normal peripheral blood mononuclear cells, reactive lymph nodes, tonsils and normal bone marrow samples showed no evidence of increased mRNA levels of CYB5B in comparison to housekeeping gene GAPDH.

Conclusions: The 21 kDa protein overexpressed in HL and aggressive NHL is identical to CYB5B. CYB5B gene expression is increased in a subset of HL and NHL cell lines tested. This is associated with CYB5B gene amplification in HL cell lines KMH2 and L428. CYB5B may be a potential target for antibody-based therapy of HL and aggressive NHL as although cytoplasmic expression is present in reactive lymphocytes, it is not expressed on the cell surface of non-neoplastic lymphocytes or bone marrow precursor cells.

Figure 1

Immunoprecipitation of the 21 kDa protein. Panel A: Western-blot analysis of immunoprecipitated 21 kDa protein (probed with anti-R23.1 MAb). Lane 1: MW = Bio-Rad Kaleidoscope molecular weight markers. Lane 2: Protein immunoprecipitated by R23.1. Lane 3: protein immunoprecipitated by R24.1. Both anti-R23.1 and anti-R24.1 MAbs precipitated a protein of ~21 kDa. Panel B: Immunoblots (probed with anti-R23.1 MAb) of subcellular fractions of cell lysates after separation of biotin-labelled surface proteins on NeutrAvidin agarose beads. Lanes 1-3: Unlabelled controls (no biotin label added to surface proteins). Lanes 4-6: Surface biotin labelled protein samples. Lane 1 and 4: Total cell lysate not further fractionated on NeutrAvidin agarose beads. Lanes 2 and 5: Proteins (cytosolic) not bound to the NeutrAvidin agarose beads. Lane 6: Cell surface biotin-labelled proteins eluted from the avidin-beads. The results show that Jurkat and L428 cells contain the protein recognised by anti-R23.1 MAb in the cytosolic fraction (Lane 5) and but only L428 cells express the protein on the cell surface (Lane 6). No protein was detected in eluted fractions from avidin beads in the absence of surface biotin labelling (Lane 3), indicating no evidence of non-specific binding of unlabelled proteins to avidin beads.

Figure 2

2D-Gel and Western-blot analysis. Panel A: Coomassie Blue staining of 2D gel. Panel B: Western-blot analysis with R23.1. The doublet in the circle is tropomyosin used as internal control. The arrow indicates the spot that reacted with R23.1 on the PVDF membrane.

Figure 3

Mass-Spectrometry data of protein isolated by 2D-gel electrophoresis. The protein was cleaved by Trypsin from C-term side of KR unless next residue is P. Matched peptides are shown in Bold Red and sequence coverage is 25%.

Figure 4

RT-PCR amplification of CYB5B. The CYB5B gene was amplified using mRNAs from L428, KMH2 and OCI/LY19 cells as templates and CYB5B-F1 and R1 primer pairs. A DNA fragment of 660 bp was detected.

Figure 5

Sequence analysis of CYB5B cDNA. Panel A: The coding region of CYB5B contains no mutation/deletion/insertion. Panel B: 150 AA translated from the CYB5B coding region.

Figure 6

Protein array identification of the molecular target of R23.1. Panel A: A PEX protein array containing approximately 16,000 bacterial clones expressing human recombinant His-tagged proteins was screened with the antibody R23.1. The antibody recognised a single clone, 9028C1242. Panel B and C: The clone 9028C1242 (full length CYB5B) and the clone 9028A0834 (expressing the COOH terminal region of CYB5B) were expressed and Western immunoblotted with both R23.1 and an anti-His antibody. R23.1 clearly recognises the approximately 20 kDa full length CYB5B, but not the COOH-region of the protein, revealing the epitope of the antibody to lie within the first 70aa.

Figure 7

Quantitative PCR of CYB5B expression in lymphoma cells lines. DLBCL (DB; OCI-LY19), HL (KMH2; L428), and ALCL (DEL; SR-786) cell lines were compared against universal human reference RNA as the calibrator. Relative expression plotted.

Figure 8

Relative mRNA expression of CYB5B in 5 Hodgkin lymphoma cell lines compared to germinal centre B cells. Fold difference is shown using normal germinal centre B cells (GC mean) as the control. Two HL cell lines L428 and KMH2 showed over 4-fold increase in CYB5B mRNA expression, while 3 HL cell lines HDLM2, L1236 and L540 decreased or similar levels of CYB5B mRNA expression (Affymetrix probe set ID 238554_at).

Figure 9

SMRT-Array CGH of HL and ALCL cell lines. High resolution karyogram of chromosome 16 showing amplification at 16q22.1 (arrow) in HL cell lines KMH2 and L428 but not ALCL cell lines DEL and SR786.

Figure 10

Immunohistochemical detection of CYB5B protein in HL clinical biopsies. An example of 2+ membrane positivity for CYB5B protein in Reed-Sternberg cells and mononuclear variants. Note prominent Golgi staining (Red arrowhead) and cell membrane detail showing a ruffled pattern (black arrow). Magnification × 600.

Figure 11

Immunohistochemical detection of CYB5B protein in HL clinical biopsies. An example of intense (3+) positivity for CYB5B on the cell membrane (black arrow) and Golgi (red arrowhead) regions of Reed-Sternberg cells and mononuclear variants. Note speckled cytoplasmic staining pattern (green arrowhead). Magnification × 600.

Figure 12

Immunohistochemical detection of CYB5B protein in reactive lymph node. Germinal centre centrocytes with cytoplasmic expression of CYB5B protein. Membrane staining is not evident. Magnification × 400.

Figure 13

Immunohistochemical detection of CYB5B protein in reactive lymph node showing dermatopathic change. Intense cytoplasmic staining for CYB5B in T-zone lymphocytes. Membrane staining is not evident. Magnification × 600.

Figure 14

Flow cytometric detection of CYB5B in Jurkat cells. Histograms show the fluorescence intensity distribution of R23.1-FITC-stained Jurkat cells, stained either for cell-surface R23.1 on viable cells (black solid line) or intracellular R23.1 within fixed cells (gray dashed line). There is no detectable surface labelling whereas intracellular labelling is evident.

Figure 15

Flow cytometric detection of CYB5B in Jurkat cells. Histograms show the fluorescence intensity distribution of R23.1-FITC-stained, fixed/permeabilized Jurkat cells blocked either with control mouse IgG (black solid line) or unlabelled R23.1 (gray dashed line). Whereas blocking with unlabelled R23.1 mAb inhibited intracellular binding of R23.1-FITC, a similar effect was not seen with control IgG.

Figure 16

Flow cytometric detection of CYB5B in bone marrow precursor cells. Bone marrow cells were gated for viable lymphocytes using forward vs. side scatter and propidium iodide exclusion, and then analysed for surface CD34-APC and R23.1-FITC expression. Numbers indicate the percentages of the total cells in the plot falling into the indicated gates. In this example, 0.4% of cells expressed dim CD34 and R23.1 and none coexpressed CD34 hi (bright) and R23.1.

Figure 17

Results of flow cytometric detection of CYB5B in bone marrow precursor cells. Vertical scatter plot of the incidence of R23.1+ cells in the CD34-hi gate. Data shown are the percentage of all CD34-hi cells which are also R23.1+ from 9 different normal bone marrow samples. Mean percentage ± s.d. = 0.3 ± 0.4.

Figure 18

Relative levels of CYB5B mRNA expression in T and B lymphocytes isolated from reactive lymph nodes and tonsils. RT-PCR for GAPDH and CYB5B from RNA extracted from lymph node (LN) and tonsil lymphocytes enriched for CD3+ T cells and CD19+ B cells showed no increase in CYB5B mRNA in comparison to GAPDH.

Figure 19

Quantification of CYB5B RT-PCR data from clinical samples. cDNA was prepared from both CD3-selected and CD19-selected cells from normal peripheral blood mononuclear cells (PBMNC), reactive tonsils, normal bone marrow (BM) and reactive lymph node samples. PCR was performed using primers to the GAPDH housekeeping gene. Bands from the GAPDH PCR were used to normalize loading for the CD3+ and CD19+ cDNA pools for each specimen. Then, PCR was performed in duplicate for CYB5B and GAPDH, using the same amount of cDNA and the same number of cycles. Bands were quantified using spot densitometry and the ratio of CYB5B/GAPDH signal was calculated. Values shown are from a typical experiment; each sample was analyzed by at least two independent PCRs. In none of the samples did the CYB5B signal exceed that of GAPDH.